Demonstration of accessing PyFLEXTRKR MCS tracking outputs

This notebook demonstrates how to access and perform simple analysis on the MCS tracking outputs from PyFLEXTRKR. It gives example code of the basic methods of loading the data using xarray, for both the tracks data and the HEALPix pixel-level data, as well as linking the two together. Some simple analysis is done.

The example largely follows that developed by Mark Muetzelfeldt, with updates to the pixel-level data in HEALPix Zarr stores.

• Authors:

• Zhe Feng (zhe.feng@pnnl.gov)
• Mark Muetzelfeldt (mark.muetzelfeldt@reading.ac.uk)

Track statistics files

The MCS tracks statistics dataset are NetCDF4 files. These are laid out as 1D, 2D and 3D arrays of data, with the main coordinate being tracks, and variable length data stored in a fixed length array (using a maximum duration of 400, equivalent to an MCS that last for 400 hr). A consequence of this is that all the data over the duration of the length of a given MCS will be NANs or similar, and so the compression ratio of files on disk is very high. Hence, even though a file for a given year is only approximately 500MB on disk, its size in memory will be far larger. Fully loading one year using xr.load_dataset uses approximately 16G of memory. It is therefore sensible to use xr.open_dataset or xr.open_mfdataset, which access the dataset's metadata but do not load its data until they are needed.

Note, the compression level of a field can be seen from xarray.Dataset: dstracks.area.encoding. The tracks dataset is compressed using compression level 4.

Pixel-level files

The pixel-level files are HEALPix format Zarr store files. Each variable is in one Zarr store. The OLR (rlut) and precipitation (pr) data are direct outputs from the models, while the MCS mask data (mcs_mask_hp(x)_) is remapped from lat/lon grid used in PyFLEXTRKR during tracking.

Using xarray to access data

xarray is a convenient way of loading NetCDF or Zarr files in Python. Fields can generally be manipulated using xarray methods or by loading the values a numpy arrays and manipulating those. This notebook requires having the following Python packages correctly installed (using e.g. conda).

# from pathlib import Path
import cftime
import cartopy.crs as ccrs
import cartopy.geodesic
import cartopy.feature as cf
import dask
from IPython.display import clear_output
import matplotlib as mpl
import matplotlib.pyplot as plt
# Use for .mp4 video:
# plt.rcParams["animation.html"] = "html5"
# Use for javascript animation:
plt.rcParams["animation.html"] = "jshtml"
import matplotlib.animation
import numpy as np
import pandas as pd
import xarray as xr

import easygems.healpix as egh

# hpdir = Path("/global/cfs/cdirs/m4549/scream-cess-healpix")
# maskdir = Path("/pscratch/sd/w/wcmca1/scream-cess-healpix/mcs_tracking_hp9/mcstracking/mcs_mask_hp9_20190901.0000_20200901.0000.zarr")
# statsdir = Path("/pscratch/sd/w/wcmca1/scream-cess-healpix/mcs_tracking_hp9/stats")
# hpdir

# Edit these based on where these are on your system.
# HEALPix data directory
hpdir = "/global/cfs/cdirs/m4549/scream-cess-healpix/"
olrdir = f"{hpdir}scream2D_hrly_rlut_hp9_v6.zarr"
prdir = f"{hpdir}scream2D_hrly_pr_hp9_v6.zarr"
# MCS data
maskdir = "/pscratch/sd/w/wcmca1/scream-cess-healpix/mcs_tracking_hp9/mcstracking/mcs_mask_hp9_20190901.0000_20200901.0000.zarr"
statsdir = "/pscratch/sd/w/wcmca1/scream-cess-healpix/mcs_tracking_hp9/stats/"

# MCS track statistics file
stats_file = f"{statsdir}mcs_tracks_final_20190901.0000_20200901.0000.nc"
stats_file
# stats_file = sorted(statsdir.glob("mcs_tracks_final_20190901.0000_20200901.0000.nc"))
# stats_file

'/pscratch/sd/w/wcmca1/scream-cess-healpix/mcs_tracking_hp9/stats/mcs_tracks_final_20190901.0000_20200901.0000.nc'

# chunks_tracks = {'tracks': 1000, 'times': 400}
chunks_tracks = None

# Open the dataset
dstracks = xr.open_dataset(
    stats_file, 
    chunks=chunks_tracks,
    mask_and_scale=True)

# # Each seperate file for each year defines its own index for tracks. Re-index with a global index.
# dstracks["tracks"] = np.arange(0, dstracks.dims["tracks"], 1, dtype=int)

dstracks.chunks

Frozen({})

# # Check if the dataset uses dask
# print(f"Dataset uses dask: {dstracks.chunks is not None}")

# # Check if variables are dask arrays
# for var_name in list(dstracks.data_vars)[:]:  # First 3 variables as example
#     print(f"Variable {var_name} is dask array: {isinstance(dstracks[var_name].data, dask.array.Array)}")

# The times have a small offset from the exact times -- e.g. 34500 ns off. Correct this.
# This mostly applies to the satellite data
def round_times_to_nearest_second(dstracks, fields):
    def remove_time_incaccuracy(t):
        # To make this an array operation, you have to use the ns version of datetime64, like so:
        return (np.round(t.astype(int) / 1e9) * 1e9).astype("datetime64[ns]")

    for field in fields:
        dstracks[field].load()
        tmask = ~np.isnan(dstracks[field].values)
        dstracks[field].values[tmask] = remove_time_incaccuracy(
            dstracks[field].values[tmask]
        )

round_times_to_nearest_second(dstracks, ['base_time', 'start_basetime', 'end_basetime'])

dstracks

<xarray.Dataset> Size: 33GB
Dimensions:                  (tracks: 52113, times: 400, mergers: 50, nmaxpf: 3)
Coordinates:
  * tracks                   (tracks) int64 417kB 0 1 2 3 ... 52110 52111 52112
  * times                    (times) int64 3kB 0 1 2 3 4 ... 395 396 397 398 399
  * mergers                  (mergers) int64 400B 0 1 2 3 4 5 ... 45 46 47 48 49
  * nmaxpf                   (nmaxpf) int64 24B 0 1 2
Data variables: (12/61)
    track_duration           (tracks) int32 208kB ...
    start_status             (tracks) float64 417kB ...
    end_status               (tracks) float64 417kB ...
    start_basetime           (tracks) datetime64[ns] 417kB 2019-09-01 ... 202...
    end_basetime             (tracks) datetime64[ns] 417kB 2019-09-01T18:00:0...
    start_split_cloudnumber  (tracks) float64 417kB ...
    ...                       ...
    pf_mcsstatus             (tracks, times) float64 167MB ...
    movement_distance        (tracks, times) float32 83MB ...
    movement_speed           (tracks, times) float32 83MB ...
    movement_theta           (tracks, times) float32 83MB ...
    movement_distance_x      (tracks, times) float32 83MB ...
    movement_distance_y      (tracks, times) float32 83MB ...
Attributes: (12/21)
    Title:                        Statistics of each MCS track
    Institution:                  Pacific Northwest National Laboratory
    Contact:                      Zhe Feng, zhe.feng@pnnl.gov
    Created_on:                   Sun Apr 13 21:03:31 2025
    startdate:                    20190901.0000
    enddate:                      20200901.0000
    ...                           ...
    MCS_PF_majoraxis_thresh:      0
    MCS_PF_duration_thresh:       4
    PF_PF_min_majoraxis_thresh:   20
    mcs_min_rainvol_thresh:       20000
    mcs_volrain_duration_thresh:  1.0
    max_speed_thresh:             50

xarray.Dataset

• Dimensions:
  • tracks: 52113
  • times: 400
  • mergers: 50
  • nmaxpf: 3
• Coordinates: (4)
  • tracks
    (tracks)
    int64
    0 1 2 3 ... 52109 52110 52111 52112

    array([    0,     1,     2, ..., 52110, 52111, 52112])

  • times
    (times)
    int64
    0 1 2 3 4 5 ... 395 396 397 398 399

    array([  0,   1,   2, ..., 397, 398, 399])

  • mergers
    (mergers)
    int64
    0 1 2 3 4 5 6 ... 44 45 46 47 48 49

    array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16, 17,
           18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35,
           36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49])

  • nmaxpf
    (nmaxpf)
    int64
    0 1 2

    array([0, 1, 2])

• Data variables: (61)
  • track_duration
    (tracks)
    int32
    ...

    long_name :

    Duration of each track

    units :

    unitless

    comments :

    Multiply by time_resolution_hour to convert to physical units

    [52113 values with dtype=int32]

  • start_status
    (tracks)
    float64
    ...

    long_name :

    Flag indicating how the track starts

    units :

    unitless

    comments :

    Refer to track_status attributes on flag value meaning

    [52113 values with dtype=float64]

  • end_status
    (tracks)
    float64
    ...

    long_name :

    Flag indicating how the track ends

    units :

    unitless

    comments :

    Refer to track_status attributes on flag value meaning

    [52113 values with dtype=float64]

  • start_basetime
    (tracks)
    datetime64[ns]
    2019-09-01 ... 2020-08-31T20:00:00

    long_name :

    Start Epoch time of each track

    array(['2019-09-01T00:00:00.000000000', '2019-09-01T00:00:00.000000000',
           '2019-09-01T00:00:00.000000000', ...,
           '2020-08-31T19:00:00.000000000', '2020-08-31T20:00:00.000000000',
           '2020-08-31T20:00:00.000000000'], dtype='datetime64[ns]')

  • end_basetime
    (tracks)
    datetime64[ns]
    2019-09-01T18:00:00 ... 2020-09-01

    long_name :

    End Epoch time of each track

    array(['2019-09-01T18:00:00.000000000', '2019-09-13T11:00:00.000000000',
           '2019-09-01T07:00:00.000000000', ...,
           '2020-09-01T00:00:00.000000000', '2020-09-01T00:00:00.000000000',
           '2020-09-01T00:00:00.000000000'], dtype='datetime64[ns]')

  • start_split_cloudnumber
    (tracks)
    float64
    ...

    long_name :

    Cloud number where this track splits from

    units :

    unitless

    [52113 values with dtype=float64]

  • end_merge_cloudnumber
    (tracks)
    float64
    ...

    long_name :

    Cloud number where this track merges with

    units :

    unitless

    [52113 values with dtype=float64]

  • pbc_flag
    (tracks)
    int32
    ...

    long_name :

    Flag indicating if position adjustment was applied due to PBC

    units :

    unitless

    comments :

    1: adjustment applied, 0: no adjustment

    [52113 values with dtype=int32]

  • base_time
    (tracks, times)
    datetime64[ns]
    2019-09-01 ... NaT

    long_name :

    Epoch time of a feature

    array([['2019-09-01T00:00:00.000000000', '2019-09-01T01:00:00.000000000',
            '2019-09-01T02:00:00.000000000', ...,
                                      'NaT',                           'NaT',
                                      'NaT'],
           ['2019-09-01T00:00:00.000000000', '2019-09-01T01:00:00.000000000',
            '2019-09-01T02:00:00.000000000', ...,
                                      'NaT',                           'NaT',
                                      'NaT'],
           ['2019-09-01T00:00:00.000000000', '2019-09-01T01:00:00.000000000',
            '2019-09-01T02:00:00.000000000', ...,
                                      'NaT',                           'NaT',
                                      'NaT'],
           ...,
           ['2020-08-31T19:00:00.000000000', '2020-08-31T20:00:00.000000000',
            '2020-08-31T21:00:00.000000000', ...,
                                      'NaT',                           'NaT',
                                      'NaT'],
           ['2020-08-31T20:00:00.000000000', '2020-08-31T21:00:00.000000000',
            '2020-08-31T22:00:00.000000000', ...,
                                      'NaT',                           'NaT',
                                      'NaT'],
           ['2020-08-31T20:00:00.000000000', '2020-08-31T21:00:00.000000000',
            '2020-08-31T22:00:00.000000000', ...,
                                      'NaT',                           'NaT',
                                      'NaT']], dtype='datetime64[ns]')

  • meanlat
    (tracks, times)
    float32
    ...

    long_name :

    Mean latitude of a feature

    units :

    degrees_north

    [20845200 values with dtype=float32]

  • meanlon
    (tracks, times)
    float32
    ...

    long_name :

    Mean longitude of a feature

    units :

    degrees_east

    [20845200 values with dtype=float32]

  • area
    (tracks, times)
    float32
    ...

    long_name :

    Area of a feature

    units :

    km^2

    [20845200 values with dtype=float32]

  • cloudnumber
    (tracks, times)
    float64
    ...

    long_name :

    Corresponding cloud number in cloudid file

    units :

    unitless

    [20845200 values with dtype=float64]

  • track_status
    (tracks, times)
    float64
    ...

    long_name :

    Flag indicating the status of a track

    units :

    unitless

    comments :

    0: Track stops; 1: Simple track continuation; 2: This is the bigger cloud in simple merger; 3: This is the bigger cloud from a simple split that stops at this time; 4: This is the bigger cloud from a split and this cloud continues to the next time; 5: This is the bigger cloud from a split that subsequently is the big cloud in a merger; 13: This cloud splits at the next time step; 15: This cloud is the bigger cloud in a merge that then splits at the next time step; 16: This is the bigger cloud in a split that then splits at the next time step; 18: Merge-split at same time (big merge, splitter, and big split); 21: This is the smaller cloud in a simple merger; 24: This is the bigger cloud of a split that is then the small cloud in a merger; 31: This is the smaller cloud in a simple split that stops; 32: This is a small split that continues onto the next time step; 33: This is a small split that then is the bigger cloud in a merger; 34: This is the small cloud in a merger that then splits at the next time step; 37: Merge-split at same time (small merge, splitter, big split); 44: This is the smaller cloud in a split that is smaller cloud in a merger at the next time step; 46: Merge-split at same time (big merge, splitter, small split); 52: This is the smaller cloud in a split that is smaller cloud in a merger at the next time step; 65: Merge-split at same time (smaller merge, splitter, small split)

    [20845200 values with dtype=float64]

  • track_interruptions
    (tracks, times)
    float64
    ...

    long_name :

    0 = full track available, good data. 1 = track starts at first file, track cut short by data availability. 2 = track ends at last file, track cut short by data availability

    units :

    unitless

    [20845200 values with dtype=float64]

  • core_area
    (tracks, times)
    float32
    ...

    long_name :

    Area of cold core

    units :

    km^2

    [20845200 values with dtype=float32]

  • cold_area
    (tracks, times)
    float32
    ...

    long_name :

    Area of cold anvil

    units :

    km^2

    [20845200 values with dtype=float32]

  • corecold_mintb
    (tracks, times)
    float32
    ...

    long_name :

    Minimum Tb in cold core + cold anvil area

    units :

    K

    [20845200 values with dtype=float32]

  • corecold_meantb
    (tracks, times)
    float32
    ...

    long_name :

    Mean Tb in cold core + cold anvil area

    units :

    K

    [20845200 values with dtype=float32]

  • core_meantb
    (tracks, times)
    float32
    ...

    long_name :

    Mean Tb in cold core area

    units :

    K

    [20845200 values with dtype=float32]

  • lat_mintb
    (tracks, times)
    float32
    ...

    long_name :

    Latitude with min Tb

    units :

    degree

    [20845200 values with dtype=float32]

  • lon_mintb
    (tracks, times)
    float32
    ...

    long_name :

    Longitude with min Tb

    units :

    degree

    [20845200 values with dtype=float32]

  • mcs_duration
    (tracks)
    int64
    ...

    long_name :

    Duration of MCS stage

    units :

    unitless

    comments :

    Multiply by time_resolution_hour to convert to physical units

    [52113 values with dtype=int64]

  • mcs_status
    (tracks, times)
    float32
    ...

    long_name :

    Flag indicating the status of MCS based on Tb. 1 = Yes, 0 = No

    units :

    unitless

    [20845200 values with dtype=float32]

  • ccs_area
    (tracks, times)
    float32
    ...

    long_name :

    Area of cold cloud shield

    units :

    km^2

    [20845200 values with dtype=float32]

  • merge_cloudnumber
    (tracks, times, mergers)
    float64
    ...

    long_name :

    Cloud numbers that merge into MCS

    units :

    unitless

    [1042260000 values with dtype=float64]

  • split_cloudnumber
    (tracks, times, mergers)
    float64
    ...

    long_name :

    Cloud numbers that split from MCS

    units :

    unitless

    [1042260000 values with dtype=float64]

  • merge_ccs_area
    (tracks, times, mergers)
    float32
    ...

    long_name :

    Cold cloud shield area that merge into MCS

    units :

    km^2

    [1042260000 values with dtype=float32]

  • split_ccs_area
    (tracks, times, mergers)
    float32
    ...

    long_name :

    Cold cloud shield area that split from MCS

    units :

    km^2

    [1042260000 values with dtype=float32]

  • pf_npf
    (tracks, times)
    float32
    ...

    long_name :

    Number of PF in the cloud

    units :

    unitless

    [20845200 values with dtype=float32]

  • pf_lon
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Mean longitude of PF

    units :

    degrees

    [62535600 values with dtype=float32]

  • pf_lat
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Mean latitude of PF

    units :

    degrees

    [62535600 values with dtype=float32]

  • pf_area
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Area of PF

    units :

    km^2

    [62535600 values with dtype=float32]

  • pf_rainrate
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Mean rain rate of PF

    units :

    mm/h

    [62535600 values with dtype=float32]

  • pf_skewness
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Rain rate skewness of PF

    units :

    unitless

    [62535600 values with dtype=float32]

  • pf_majoraxis
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Major axis length of PF

    units :

    km

    [62535600 values with dtype=float32]

  • pf_minoraxis
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Minor axis length of PF

    units :

    km

    [62535600 values with dtype=float32]

  • pf_aspectratio
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Aspect ratio (major axis / minor axis) of PF

    units :

    unitless

    [62535600 values with dtype=float32]

  • pf_orientation
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Orientation of major axis of PF

    units :

    degrees

    [62535600 values with dtype=float32]

  • pf_perimeter
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Perimeter of PF

    units :

    km

    [62535600 values with dtype=float32]

  • pf_eccentricity
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Eccentricity of PF

    units :

    unitless

    [62535600 values with dtype=float32]

  • pf_lon_centroid
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Centroid longitude of PF

    units :

    degrees

    [62535600 values with dtype=float32]

  • pf_lat_centroid
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Centroid latitude of PF

    units :

    degrees

    [62535600 values with dtype=float32]

  • pf_lon_weightedcentroid
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Weighted centroid longitude of PF

    units :

    degrees

    [62535600 values with dtype=float32]

  • pf_lat_weightedcentroid
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Weighted centroid latitude of PF

    units :

    degrees

    [62535600 values with dtype=float32]

  • pf_lon_maxrainrate
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Longitude with max rain rate

    units :

    degree

    [62535600 values with dtype=float32]

  • pf_lat_maxrainrate
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Latitude with max rain rate

    units :

    degree

    [62535600 values with dtype=float32]

  • pf_maxrainrate
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Max rain rate of PF

    units :

    mm/h

    [62535600 values with dtype=float32]

  • pf_accumrain
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Accumulate precipitation of PF

    units :

    mm/h

    [62535600 values with dtype=float32]

  • pf_accumrainheavy
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Accumulated heavy precipitation of PF

    units :

    mm/h

    heavy_rainrate_threshold :

    10.0

    [62535600 values with dtype=float32]

  • pf_landfrac
    (tracks, times)
    float32
    ...

    long_name :

    Fraction of PF over land

    units :

    fraction

    [20845200 values with dtype=float32]

  • total_rain
    (tracks, times)
    float32
    ...

    long_name :

    Total precipitation under cold cloud shield (all rainfall included)

    units :

    mm/h

    [20845200 values with dtype=float32]

  • total_heavyrain
    (tracks, times)
    float32
    ...

    long_name :

    Total heavy precipitation under cold cloud shield

    units :

    mm/h

    heavy_rainrate_threshold :

    10.0

    [20845200 values with dtype=float32]

  • rainrate_heavyrain
    (tracks, times)
    float32
    ...

    long_name :

    Mean heavy rain rate under cold cloud shield

    units :

    mm/h

    heavy_rainrate_threshold :

    10.0

    [20845200 values with dtype=float32]

  • pf_lifetime
    (tracks)
    float32
    ...

    long_name :

    MCS lifetime when a significant PF is present

    units :

    hour

    [52113 values with dtype=float32]

  • pf_mcsstatus
    (tracks, times)
    float64
    ...

    long_name :

    Flag indicating the status of MCS based on PF. 1 = Yes, 0 = No

    units :

    unitless

    [20845200 values with dtype=float64]

  • movement_distance
    (tracks, times)
    float32
    ...

    long_name :

    Movement distance along angle movement_theta

    units :

    km

    lag :

    1

    comments :

    This is the total movement along the angle theta between lag estimates.

    [20845200 values with dtype=float32]

  • movement_speed
    (tracks, times)
    float32
    ...

    long_name :

    Movement speed along angle movement_theta

    units :

    m/s

    [20845200 values with dtype=float32]

  • movement_theta
    (tracks, times)
    float32
    ...

    long_name :

    Movement direction

    units :

    degrees

    comments :

    0:northward, 90:eastward, 180:southward, 270:westward.

    [20845200 values with dtype=float32]

  • movement_distance_x
    (tracks, times)
    float32
    ...

    long_name :

    East-West component of movement distance

    units :

    km

    [20845200 values with dtype=float32]

  • movement_distance_y
    (tracks, times)
    float32
    ...

    long_name :

    North-South component of movement distance

    units :

    km

    [20845200 values with dtype=float32]

• Indexes: (4)
  • tracks
    PandasIndex

    PandasIndex(Index([    0,     1,     2,     3,     4,     5,     6,     7,     8,     9,
           ...
           52103, 52104, 52105, 52106, 52107, 52108, 52109, 52110, 52111, 52112],
          dtype='int64', name='tracks', length=52113))

  • times
    PandasIndex

    PandasIndex(Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
           ...
           390, 391, 392, 393, 394, 395, 396, 397, 398, 399],
          dtype='int64', name='times', length=400))

  • mergers
    PandasIndex

    PandasIndex(Index([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16, 17,
           18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35,
           36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49],
          dtype='int64', name='mergers'))

  • nmaxpf
    PandasIndex

    PandasIndex(Index([0, 1, 2], dtype='int64', name='nmaxpf'))

• Attributes: (21)

  Title :

  Statistics of each MCS track

  Institution :

  Pacific Northwest National Laboratory

  Contact :

  Zhe Feng, zhe.feng@pnnl.gov

  Created_on :

  Sun Apr 13 21:03:31 2025

  startdate :

  20190901.0000

  enddate :

  20200901.0000

  timegap_hour :

  3.1

  time_resolution_hour :

  1.0

  pixel_radius_km :

  10.0

  MCS_duration_hr :

  4

  MCS_area_km^2 :

  40000

  nmaxpf :

  3

  PF_rainrate_thresh :

  2.0

  heavy_rainrate_thresh :

  10.0

  landfrac_thresh :

  [ 0 90]

  MCS_PF_majoraxis_thresh :

  0

  MCS_PF_duration_thresh :

  4

  PF_PF_min_majoraxis_thresh :

  20

  mcs_min_rainvol_thresh :

  20000

  mcs_volrain_duration_thresh :

  1.0

  max_speed_thresh :

  50

# Note, 'NaT' means nan (track does not exist at those times)
dstracks.base_time

<xarray.DataArray 'base_time' (tracks: 52113, times: 400)> Size: 167MB
array([['2019-09-01T00:00:00.000000000', '2019-09-01T01:00:00.000000000',
        '2019-09-01T02:00:00.000000000', ...,
                                  'NaT',                           'NaT',
                                  'NaT'],
       ['2019-09-01T00:00:00.000000000', '2019-09-01T01:00:00.000000000',
        '2019-09-01T02:00:00.000000000', ...,
                                  'NaT',                           'NaT',
                                  'NaT'],
       ['2019-09-01T00:00:00.000000000', '2019-09-01T01:00:00.000000000',
        '2019-09-01T02:00:00.000000000', ...,
                                  'NaT',                           'NaT',
                                  'NaT'],
       ...,
       ['2020-08-31T19:00:00.000000000', '2020-08-31T20:00:00.000000000',
        '2020-08-31T21:00:00.000000000', ...,
                                  'NaT',                           'NaT',
                                  'NaT'],
       ['2020-08-31T20:00:00.000000000', '2020-08-31T21:00:00.000000000',
        '2020-08-31T22:00:00.000000000', ...,
                                  'NaT',                           'NaT',
                                  'NaT'],
       ['2020-08-31T20:00:00.000000000', '2020-08-31T21:00:00.000000000',
        '2020-08-31T22:00:00.000000000', ...,
                                  'NaT',                           'NaT',
                                  'NaT']], dtype='datetime64[ns]')
Coordinates:
  * tracks   (tracks) int64 417kB 0 1 2 3 4 5 ... 52108 52109 52110 52111 52112
  * times    (times) int64 3kB 0 1 2 3 4 5 6 7 ... 393 394 395 396 397 398 399
Attributes:
    long_name:  Epoch time of a feature

xarray.DataArray

'base_time'

• tracks: 52113
• times: 400

• 2019-09-01 2019-09-01T01:00:00 2019-09-01T02:00:00 ... NaT NaT NaT

  array([['2019-09-01T00:00:00.000000000', '2019-09-01T01:00:00.000000000',
          '2019-09-01T02:00:00.000000000', ...,
                                    'NaT',                           'NaT',
                                    'NaT'],
         ['2019-09-01T00:00:00.000000000', '2019-09-01T01:00:00.000000000',
          '2019-09-01T02:00:00.000000000', ...,
                                    'NaT',                           'NaT',
                                    'NaT'],
         ['2019-09-01T00:00:00.000000000', '2019-09-01T01:00:00.000000000',
          '2019-09-01T02:00:00.000000000', ...,
                                    'NaT',                           'NaT',
                                    'NaT'],
         ...,
         ['2020-08-31T19:00:00.000000000', '2020-08-31T20:00:00.000000000',
          '2020-08-31T21:00:00.000000000', ...,
                                    'NaT',                           'NaT',
                                    'NaT'],
         ['2020-08-31T20:00:00.000000000', '2020-08-31T21:00:00.000000000',
          '2020-08-31T22:00:00.000000000', ...,
                                    'NaT',                           'NaT',
                                    'NaT'],
         ['2020-08-31T20:00:00.000000000', '2020-08-31T21:00:00.000000000',
          '2020-08-31T22:00:00.000000000', ...,
                                    'NaT',                           'NaT',
                                    'NaT']], dtype='datetime64[ns]')

• Coordinates: (2)
  • tracks
    (tracks)
    int64
    0 1 2 3 ... 52109 52110 52111 52112

    array([    0,     1,     2, ..., 52110, 52111, 52112])

  • times
    (times)
    int64
    0 1 2 3 4 5 ... 395 396 397 398 399

    array([  0,   1,   2, ..., 397, 398, 399])

• Indexes: (2)
  • tracks
    PandasIndex

    PandasIndex(Index([    0,     1,     2,     3,     4,     5,     6,     7,     8,     9,
           ...
           52103, 52104, 52105, 52106, 52107, 52108, 52109, 52110, 52111, 52112],
          dtype='int64', name='tracks', length=52113))

  • times
    PandasIndex

    PandasIndex(Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
           ...
           390, 391, 392, 393, 394, 395, 396, 397, 398, 399],
          dtype='int64', name='times', length=400))

• Attributes: (1)

  long_name :

  Epoch time of a feature

# Values can be accessed as a `numpy` array:
dstracks.ccs_area.values

array([[1209400.,  689200., 1902000., ...,      nan,      nan,      nan],
       [ 704400.,  667300.,  744000., ...,      nan,      nan,      nan],
       [ 588200.,  495000.,  287000., ...,      nan,      nan,      nan],
       ...,
       [   3300.,   25300.,   52500., ...,      nan,      nan,      nan],
       [ 381300.,  402900.,  280000., ...,      nan,      nan,      nan],
       [  53000.,   50400.,   54300., ...,      nan,      nan,      nan]],
      dtype=float32)

# Compression level info:
dstracks.ccs_area.encoding

{'dtype': dtype('float32'),
 'zlib': True,
 'szip': False,
 'zstd': False,
 'bzip2': False,
 'blosc': False,
 'shuffle': True,
 'complevel': 4,
 'fletcher32': False,
 'contiguous': False,
 'chunksizes': (1, 400),
 'preferred_chunks': {'tracks': 1, 'times': 400},
 'source': '/pscratch/sd/w/wcmca1/scream-cess-healpix/mcs_tracking_hp9/stats/mcs_tracks_final_20190901.0000_20200901.0000.nc',
 'original_shape': (52113, 400),
 '_FillValue': np.float32(nan)}

Selecting tracks

# A single track can be selected from its track number:
track = dstracks.sel(tracks=234)
track

<xarray.Dataset> Size: 631kB
Dimensions:                  (times: 400, mergers: 50, nmaxpf: 3)
Coordinates:
    tracks                   int64 8B 234
  * times                    (times) int64 3kB 0 1 2 3 4 ... 395 396 397 398 399
  * mergers                  (mergers) int64 400B 0 1 2 3 4 5 ... 45 46 47 48 49
  * nmaxpf                   (nmaxpf) int64 24B 0 1 2
Data variables: (12/61)
    track_duration           int32 4B ...
    start_status             float64 8B ...
    end_status               float64 8B ...
    start_basetime           datetime64[ns] 8B 2019-09-02T08:00:00
    end_basetime             datetime64[ns] 8B 2019-09-03T23:00:00
    start_split_cloudnumber  float64 8B ...
    ...                       ...
    pf_mcsstatus             (times) float64 3kB ...
    movement_distance        (times) float32 2kB ...
    movement_speed           (times) float32 2kB ...
    movement_theta           (times) float32 2kB ...
    movement_distance_x      (times) float32 2kB ...
    movement_distance_y      (times) float32 2kB ...
Attributes: (12/21)
    Title:                        Statistics of each MCS track
    Institution:                  Pacific Northwest National Laboratory
    Contact:                      Zhe Feng, zhe.feng@pnnl.gov
    Created_on:                   Sun Apr 13 21:03:31 2025
    startdate:                    20190901.0000
    enddate:                      20200901.0000
    ...                           ...
    MCS_PF_majoraxis_thresh:      0
    MCS_PF_duration_thresh:       4
    PF_PF_min_majoraxis_thresh:   20
    mcs_min_rainvol_thresh:       20000
    mcs_volrain_duration_thresh:  1.0
    max_speed_thresh:             50

xarray.Dataset

• Dimensions:
  • times: 400
  • mergers: 50
  • nmaxpf: 3
• Coordinates: (4)
  • tracks
    ()
    int64
    234

    array(234)

  • times
    (times)
    int64
    0 1 2 3 4 5 ... 395 396 397 398 399

    array([  0,   1,   2, ..., 397, 398, 399])

  • mergers
    (mergers)
    int64
    0 1 2 3 4 5 6 ... 44 45 46 47 48 49

    array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16, 17,
           18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35,
           36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49])

  • nmaxpf
    (nmaxpf)
    int64
    0 1 2

    array([0, 1, 2])

• Data variables: (61)
  • track_duration
    ()
    int32
    ...

    long_name :

    Duration of each track

    units :

    unitless

    comments :

    Multiply by time_resolution_hour to convert to physical units

    [1 values with dtype=int32]

  • start_status
    ()
    float64
    ...

    long_name :

    Flag indicating how the track starts

    units :

    unitless

    comments :

    Refer to track_status attributes on flag value meaning

    [1 values with dtype=float64]

  • end_status
    ()
    float64
    ...

    long_name :

    Flag indicating how the track ends

    units :

    unitless

    comments :

    Refer to track_status attributes on flag value meaning

    [1 values with dtype=float64]

  • start_basetime
    ()
    datetime64[ns]
    2019-09-02T08:00:00

    long_name :

    Start Epoch time of each track

    array('2019-09-02T08:00:00.000000000', dtype='datetime64[ns]')

  • end_basetime
    ()
    datetime64[ns]
    2019-09-03T23:00:00

    long_name :

    End Epoch time of each track

    array('2019-09-03T23:00:00.000000000', dtype='datetime64[ns]')

  • start_split_cloudnumber
    ()
    float64
    ...

    long_name :

    Cloud number where this track splits from

    units :

    unitless

    [1 values with dtype=float64]

  • end_merge_cloudnumber
    ()
    float64
    ...

    long_name :

    Cloud number where this track merges with

    units :

    unitless

    [1 values with dtype=float64]

  • pbc_flag
    ()
    int32
    ...

    long_name :

    Flag indicating if position adjustment was applied due to PBC

    units :

    unitless

    comments :

    1: adjustment applied, 0: no adjustment

    [1 values with dtype=int32]

  • base_time
    (times)
    datetime64[ns]
    2019-09-02T08:00:00 ... NaT

    long_name :

    Epoch time of a feature

    array(['2019-09-02T08:00:00.000000000', '2019-09-02T09:00:00.000000000',
           '2019-09-02T10:00:00.000000000', '2019-09-02T11:00:00.000000000',
           '2019-09-02T12:00:00.000000000', '2019-09-02T13:00:00.000000000',
           '2019-09-02T14:00:00.000000000', '2019-09-02T15:00:00.000000000',
           '2019-09-02T16:00:00.000000000', '2019-09-02T17:00:00.000000000',
           '2019-09-02T18:00:00.000000000', '2019-09-02T19:00:00.000000000',
           '2019-09-02T20:00:00.000000000', '2019-09-02T21:00:00.000000000',
           '2019-09-02T22:00:00.000000000', '2019-09-02T23:00:00.000000000',
           '2019-09-03T00:00:00.000000000', '2019-09-03T01:00:00.000000000',
           '2019-09-03T02:00:00.000000000', '2019-09-03T03:00:00.000000000',
           '2019-09-03T04:00:00.000000000', '2019-09-03T05:00:00.000000000',
           '2019-09-03T06:00:00.000000000', '2019-09-03T07:00:00.000000000',
           '2019-09-03T08:00:00.000000000', '2019-09-03T09:00:00.000000000',
           '2019-09-03T10:00:00.000000000', '2019-09-03T11:00:00.000000000',
           '2019-09-03T12:00:00.000000000', '2019-09-03T13:00:00.000000000',
           '2019-09-03T14:00:00.000000000', '2019-09-03T15:00:00.000000000',
           '2019-09-03T16:00:00.000000000', '2019-09-03T17:00:00.000000000',
           '2019-09-03T18:00:00.000000000', '2019-09-03T19:00:00.000000000',
           '2019-09-03T20:00:00.000000000', '2019-09-03T21:00:00.000000000',
           '2019-09-03T22:00:00.000000000', '2019-09-03T23:00:00.000000000',
    ...
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT'],
          dtype='datetime64[ns]')

  • meanlat
    (times)
    float32
    ...

    long_name :

    Mean latitude of a feature

    units :

    degrees_north

    [400 values with dtype=float32]

  • meanlon
    (times)
    float32
    ...

    long_name :

    Mean longitude of a feature

    units :

    degrees_east

    [400 values with dtype=float32]

  • area
    (times)
    float32
    ...

    long_name :

    Area of a feature

    units :

    km^2

    [400 values with dtype=float32]

  • cloudnumber
    (times)
    float64
    ...

    long_name :

    Corresponding cloud number in cloudid file

    units :

    unitless

    [400 values with dtype=float64]

  • track_status
    (times)
    float64
    ...

    long_name :

    Flag indicating the status of a track

    units :

    unitless

    comments :

    0: Track stops; 1: Simple track continuation; 2: This is the bigger cloud in simple merger; 3: This is the bigger cloud from a simple split that stops at this time; 4: This is the bigger cloud from a split and this cloud continues to the next time; 5: This is the bigger cloud from a split that subsequently is the big cloud in a merger; 13: This cloud splits at the next time step; 15: This cloud is the bigger cloud in a merge that then splits at the next time step; 16: This is the bigger cloud in a split that then splits at the next time step; 18: Merge-split at same time (big merge, splitter, and big split); 21: This is the smaller cloud in a simple merger; 24: This is the bigger cloud of a split that is then the small cloud in a merger; 31: This is the smaller cloud in a simple split that stops; 32: This is a small split that continues onto the next time step; 33: This is a small split that then is the bigger cloud in a merger; 34: This is the small cloud in a merger that then splits at the next time step; 37: Merge-split at same time (small merge, splitter, big split); 44: This is the smaller cloud in a split that is smaller cloud in a merger at the next time step; 46: Merge-split at same time (big merge, splitter, small split); 52: This is the smaller cloud in a split that is smaller cloud in a merger at the next time step; 65: Merge-split at same time (smaller merge, splitter, small split)

    [400 values with dtype=float64]

  • track_interruptions
    (times)
    float64
    ...

    long_name :

    0 = full track available, good data. 1 = track starts at first file, track cut short by data availability. 2 = track ends at last file, track cut short by data availability

    units :

    unitless

    [400 values with dtype=float64]

  • core_area
    (times)
    float32
    ...

    long_name :

    Area of cold core

    units :

    km^2

    [400 values with dtype=float32]

  • cold_area
    (times)
    float32
    ...

    long_name :

    Area of cold anvil

    units :

    km^2

    [400 values with dtype=float32]

  • corecold_mintb
    (times)
    float32
    ...

    long_name :

    Minimum Tb in cold core + cold anvil area

    units :

    K

    [400 values with dtype=float32]

  • corecold_meantb
    (times)
    float32
    ...

    long_name :

    Mean Tb in cold core + cold anvil area

    units :

    K

    [400 values with dtype=float32]

  • core_meantb
    (times)
    float32
    ...

    long_name :

    Mean Tb in cold core area

    units :

    K

    [400 values with dtype=float32]

  • lat_mintb
    (times)
    float32
    ...

    long_name :

    Latitude with min Tb

    units :

    degree

    [400 values with dtype=float32]

  • lon_mintb
    (times)
    float32
    ...

    long_name :

    Longitude with min Tb

    units :

    degree

    [400 values with dtype=float32]

  • mcs_duration
    ()
    int64
    ...

    long_name :

    Duration of MCS stage

    units :

    unitless

    comments :

    Multiply by time_resolution_hour to convert to physical units

    [1 values with dtype=int64]

  • mcs_status
    (times)
    float32
    ...

    long_name :

    Flag indicating the status of MCS based on Tb. 1 = Yes, 0 = No

    units :

    unitless

    [400 values with dtype=float32]

  • ccs_area
    (times)
    float32
    1.773e+05 1.73e+05 ... nan nan

    long_name :

    Area of cold cloud shield

    units :

    km^2

    array([177300., 173000., 150300., ...,     nan,     nan,     nan],
          dtype=float32)

  • merge_cloudnumber
    (times, mergers)
    float64
    ...

    long_name :

    Cloud numbers that merge into MCS

    units :

    unitless

    [20000 values with dtype=float64]

  • split_cloudnumber
    (times, mergers)
    float64
    ...

    long_name :

    Cloud numbers that split from MCS

    units :

    unitless

    [20000 values with dtype=float64]

  • merge_ccs_area
    (times, mergers)
    float32
    ...

    long_name :

    Cold cloud shield area that merge into MCS

    units :

    km^2

    [20000 values with dtype=float32]

  • split_ccs_area
    (times, mergers)
    float32
    ...

    long_name :

    Cold cloud shield area that split from MCS

    units :

    km^2

    [20000 values with dtype=float32]

  • pf_npf
    (times)
    float32
    ...

    long_name :

    Number of PF in the cloud

    units :

    unitless

    [400 values with dtype=float32]

  • pf_lon
    (times, nmaxpf)
    float32
    ...

    long_name :

    Mean longitude of PF

    units :

    degrees

    [1200 values with dtype=float32]

  • pf_lat
    (times, nmaxpf)
    float32
    ...

    long_name :

    Mean latitude of PF

    units :

    degrees

    [1200 values with dtype=float32]

  • pf_area
    (times, nmaxpf)
    float32
    ...

    long_name :

    Area of PF

    units :

    km^2

    [1200 values with dtype=float32]

  • pf_rainrate
    (times, nmaxpf)
    float32
    ...

    long_name :

    Mean rain rate of PF

    units :

    mm/h

    [1200 values with dtype=float32]

  • pf_skewness
    (times, nmaxpf)
    float32
    ...

    long_name :

    Rain rate skewness of PF

    units :

    unitless

    [1200 values with dtype=float32]

  • pf_majoraxis
    (times, nmaxpf)
    float32
    ...

    long_name :

    Major axis length of PF

    units :

    km

    [1200 values with dtype=float32]

  • pf_minoraxis
    (times, nmaxpf)
    float32
    ...

    long_name :

    Minor axis length of PF

    units :

    km

    [1200 values with dtype=float32]

  • pf_aspectratio
    (times, nmaxpf)
    float32
    ...

    long_name :

    Aspect ratio (major axis / minor axis) of PF

    units :

    unitless

    [1200 values with dtype=float32]

  • pf_orientation
    (times, nmaxpf)
    float32
    ...

    long_name :

    Orientation of major axis of PF

    units :

    degrees

    [1200 values with dtype=float32]

  • pf_perimeter
    (times, nmaxpf)
    float32
    ...

    long_name :

    Perimeter of PF

    units :

    km

    [1200 values with dtype=float32]

  • pf_eccentricity
    (times, nmaxpf)
    float32
    ...

    long_name :

    Eccentricity of PF

    units :

    unitless

    [1200 values with dtype=float32]

  • pf_lon_centroid
    (times, nmaxpf)
    float32
    ...

    long_name :

    Centroid longitude of PF

    units :

    degrees

    [1200 values with dtype=float32]

  • pf_lat_centroid
    (times, nmaxpf)
    float32
    ...

    long_name :

    Centroid latitude of PF

    units :

    degrees

    [1200 values with dtype=float32]

  • pf_lon_weightedcentroid
    (times, nmaxpf)
    float32
    ...

    long_name :

    Weighted centroid longitude of PF

    units :

    degrees

    [1200 values with dtype=float32]

  • pf_lat_weightedcentroid
    (times, nmaxpf)
    float32
    ...

    long_name :

    Weighted centroid latitude of PF

    units :

    degrees

    [1200 values with dtype=float32]

  • pf_lon_maxrainrate
    (times, nmaxpf)
    float32
    ...

    long_name :

    Longitude with max rain rate

    units :

    degree

    [1200 values with dtype=float32]

  • pf_lat_maxrainrate
    (times, nmaxpf)
    float32
    ...

    long_name :

    Latitude with max rain rate

    units :

    degree

    [1200 values with dtype=float32]

  • pf_maxrainrate
    (times, nmaxpf)
    float32
    ...

    long_name :

    Max rain rate of PF

    units :

    mm/h

    [1200 values with dtype=float32]

  • pf_accumrain
    (times, nmaxpf)
    float32
    ...

    long_name :

    Accumulate precipitation of PF

    units :

    mm/h

    [1200 values with dtype=float32]

  • pf_accumrainheavy
    (times, nmaxpf)
    float32
    ...

    long_name :

    Accumulated heavy precipitation of PF

    units :

    mm/h

    heavy_rainrate_threshold :

    10.0

    [1200 values with dtype=float32]

  • pf_landfrac
    (times)
    float32
    ...

    long_name :

    Fraction of PF over land

    units :

    fraction

    [400 values with dtype=float32]

  • total_rain
    (times)
    float32
    ...

    long_name :

    Total precipitation under cold cloud shield (all rainfall included)

    units :

    mm/h

    [400 values with dtype=float32]

  • total_heavyrain
    (times)
    float32
    ...

    long_name :

    Total heavy precipitation under cold cloud shield

    units :

    mm/h

    heavy_rainrate_threshold :

    10.0

    [400 values with dtype=float32]

  • rainrate_heavyrain
    (times)
    float32
    ...

    long_name :

    Mean heavy rain rate under cold cloud shield

    units :

    mm/h

    heavy_rainrate_threshold :

    10.0

    [400 values with dtype=float32]

  • pf_lifetime
    ()
    float32
    ...

    long_name :

    MCS lifetime when a significant PF is present

    units :

    hour

    [1 values with dtype=float32]

  • pf_mcsstatus
    (times)
    float64
    ...

    long_name :

    Flag indicating the status of MCS based on PF. 1 = Yes, 0 = No

    units :

    unitless

    [400 values with dtype=float64]

  • movement_distance
    (times)
    float32
    ...

    long_name :

    Movement distance along angle movement_theta

    units :

    km

    lag :

    1

    comments :

    This is the total movement along the angle theta between lag estimates.

    [400 values with dtype=float32]

  • movement_speed
    (times)
    float32
    ...

    long_name :

    Movement speed along angle movement_theta

    units :

    m/s

    [400 values with dtype=float32]

  • movement_theta
    (times)
    float32
    ...

    long_name :

    Movement direction

    units :

    degrees

    comments :

    0:northward, 90:eastward, 180:southward, 270:westward.

    [400 values with dtype=float32]

  • movement_distance_x
    (times)
    float32
    ...

    long_name :

    East-West component of movement distance

    units :

    km

    [400 values with dtype=float32]

  • movement_distance_y
    (times)
    float32
    ...

    long_name :

    North-South component of movement distance

    units :

    km

    [400 values with dtype=float32]

• Indexes: (3)
  • times
    PandasIndex

    PandasIndex(Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
           ...
           390, 391, 392, 393, 394, 395, 396, 397, 398, 399],
          dtype='int64', name='times', length=400))

  • mergers
    PandasIndex

    PandasIndex(Index([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16, 17,
           18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35,
           36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49],
          dtype='int64', name='mergers'))

  • nmaxpf
    PandasIndex

    PandasIndex(Index([0, 1, 2], dtype='int64', name='nmaxpf'))

• Attributes: (21)

  Title :

  Statistics of each MCS track

  Institution :

  Pacific Northwest National Laboratory

  Contact :

  Zhe Feng, zhe.feng@pnnl.gov

  Created_on :

  Sun Apr 13 21:03:31 2025

  startdate :

  20190901.0000

  enddate :

  20200901.0000

  timegap_hour :

  3.1

  time_resolution_hour :

  1.0

  pixel_radius_km :

  10.0

  MCS_duration_hr :

  4

  MCS_area_km^2 :

  40000

  nmaxpf :

  3

  PF_rainrate_thresh :

  2.0

  heavy_rainrate_thresh :

  10.0

  landfrac_thresh :

  [ 0 90]

  MCS_PF_majoraxis_thresh :

  0

  MCS_PF_duration_thresh :

  4

  PF_PF_min_majoraxis_thresh :

  20

  mcs_min_rainvol_thresh :

  20000

  mcs_volrain_duration_thresh :

  1.0

  max_speed_thresh :

  50

# You might want to select tracks based on e.g. the time at which they are active:
datetime = pd.Timestamp('2020-02-01 12:00').to_numpy()
# `isel` selects on index. The expression on the RHS collapses the 2D field of `base_time` into a 
# 1D boolean field that is true if *any* base_time for a given track matches `datetime`.
dstracks_at_time = dstracks.isel(
     tracks=(dstracks.base_time.values == datetime).any(axis=1)
)
dstracks_at_time

<xarray.Dataset> Size: 63MB
Dimensions:                  (tracks: 100, times: 400, mergers: 50, nmaxpf: 3)
Coordinates:
  * tracks                   (tracks) int64 800B 20832 21069 ... 21446 21447
  * times                    (times) int64 3kB 0 1 2 3 4 ... 395 396 397 398 399
  * mergers                  (mergers) int64 400B 0 1 2 3 4 5 ... 45 46 47 48 49
  * nmaxpf                   (nmaxpf) int64 24B 0 1 2
Data variables: (12/61)
    track_duration           (tracks) int32 400B ...
    start_status             (tracks) float64 800B ...
    end_status               (tracks) float64 800B ...
    start_basetime           (tracks) datetime64[ns] 800B 2020-01-27T21:00:00...
    end_basetime             (tracks) datetime64[ns] 800B 2020-02-01T19:00:00...
    start_split_cloudnumber  (tracks) float64 800B ...
    ...                       ...
    pf_mcsstatus             (tracks, times) float64 320kB ...
    movement_distance        (tracks, times) float32 160kB ...
    movement_speed           (tracks, times) float32 160kB ...
    movement_theta           (tracks, times) float32 160kB ...
    movement_distance_x      (tracks, times) float32 160kB ...
    movement_distance_y      (tracks, times) float32 160kB ...
Attributes: (12/21)
    Title:                        Statistics of each MCS track
    Institution:                  Pacific Northwest National Laboratory
    Contact:                      Zhe Feng, zhe.feng@pnnl.gov
    Created_on:                   Sun Apr 13 21:03:31 2025
    startdate:                    20190901.0000
    enddate:                      20200901.0000
    ...                           ...
    MCS_PF_majoraxis_thresh:      0
    MCS_PF_duration_thresh:       4
    PF_PF_min_majoraxis_thresh:   20
    mcs_min_rainvol_thresh:       20000
    mcs_volrain_duration_thresh:  1.0
    max_speed_thresh:             50

xarray.Dataset

• Dimensions:
  • tracks: 100
  • times: 400
  • mergers: 50
  • nmaxpf: 3
• Coordinates: (4)
  • tracks
    (tracks)
    int64
    20832 21069 21085 ... 21446 21447

    array([20832, 21069, 21085, 21099, 21166, 21199, 21205, 21210, 21217, 21221,
           21231, 21248, 21260, 21271, 21290, 21294, 21300, 21304, 21319, 21328,
           21329, 21336, 21337, 21341, 21344, 21345, 21347, 21348, 21352, 21356,
           21358, 21360, 21363, 21364, 21365, 21367, 21370, 21372, 21373, 21374,
           21375, 21376, 21380, 21381, 21382, 21383, 21386, 21393, 21394, 21395,
           21396, 21398, 21399, 21400, 21401, 21402, 21403, 21404, 21405, 21406,
           21407, 21408, 21409, 21410, 21411, 21412, 21413, 21415, 21416, 21417,
           21418, 21419, 21420, 21421, 21422, 21423, 21424, 21425, 21426, 21427,
           21428, 21429, 21430, 21431, 21432, 21433, 21434, 21435, 21436, 21437,
           21438, 21439, 21440, 21441, 21442, 21443, 21444, 21445, 21446, 21447])

  • times
    (times)
    int64
    0 1 2 3 4 5 ... 395 396 397 398 399

    array([  0,   1,   2, ..., 397, 398, 399])

  • mergers
    (mergers)
    int64
    0 1 2 3 4 5 6 ... 44 45 46 47 48 49

    array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16, 17,
           18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35,
           36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49])

  • nmaxpf
    (nmaxpf)
    int64
    0 1 2

    array([0, 1, 2])

• Data variables: (61)
  • track_duration
    (tracks)
    int32
    ...

    long_name :

    Duration of each track

    units :

    unitless

    comments :

    Multiply by time_resolution_hour to convert to physical units

    [100 values with dtype=int32]

  • start_status
    (tracks)
    float64
    ...

    long_name :

    Flag indicating how the track starts

    units :

    unitless

    comments :

    Refer to track_status attributes on flag value meaning

    [100 values with dtype=float64]

  • end_status
    (tracks)
    float64
    ...

    long_name :

    Flag indicating how the track ends

    units :

    unitless

    comments :

    Refer to track_status attributes on flag value meaning

    [100 values with dtype=float64]

  • start_basetime
    (tracks)
    datetime64[ns]
    2020-01-27T21:00:00 ... 2020-02-...

    long_name :

    Start Epoch time of each track

    array(['2020-01-27T21:00:00.000000000', '2020-01-29T20:00:00.000000000',
           '2020-01-29T22:00:00.000000000', '2020-01-30T02:00:00.000000000',
           '2020-01-30T14:00:00.000000000', '2020-01-30T18:00:00.000000000',
           '2020-01-30T19:00:00.000000000', '2020-01-30T20:00:00.000000000',
           '2020-01-30T21:00:00.000000000', '2020-01-30T22:00:00.000000000',
           '2020-01-30T23:00:00.000000000', '2020-01-31T02:00:00.000000000',
           '2020-01-31T04:00:00.000000000', '2020-01-31T05:00:00.000000000',
           '2020-01-31T07:00:00.000000000', '2020-01-31T08:00:00.000000000',
           '2020-01-31T09:00:00.000000000', '2020-01-31T10:00:00.000000000',
           '2020-01-31T12:00:00.000000000', '2020-01-31T14:00:00.000000000',
           '2020-01-31T14:00:00.000000000', '2020-01-31T16:00:00.000000000',
           '2020-01-31T16:00:00.000000000', '2020-01-31T17:00:00.000000000',
           '2020-01-31T18:00:00.000000000', '2020-01-31T18:00:00.000000000',
           '2020-01-31T18:00:00.000000000', '2020-01-31T19:00:00.000000000',
           '2020-01-31T19:00:00.000000000', '2020-01-31T20:00:00.000000000',
           '2020-01-31T20:00:00.000000000', '2020-01-31T20:00:00.000000000',
           '2020-01-31T21:00:00.000000000', '2020-01-31T21:00:00.000000000',
           '2020-01-31T22:00:00.000000000', '2020-01-31T22:00:00.000000000',
           '2020-01-31T23:00:00.000000000', '2020-01-31T23:00:00.000000000',
           '2020-01-31T23:00:00.000000000', '2020-01-31T23:00:00.000000000',
    ...
           '2020-02-01T04:00:00.000000000', '2020-02-01T05:00:00.000000000',
           '2020-02-01T05:00:00.000000000', '2020-02-01T06:00:00.000000000',
           '2020-02-01T06:00:00.000000000', '2020-02-01T06:00:00.000000000',
           '2020-02-01T07:00:00.000000000', '2020-02-01T07:00:00.000000000',
           '2020-02-01T07:00:00.000000000', '2020-02-01T07:00:00.000000000',
           '2020-02-01T07:00:00.000000000', '2020-02-01T08:00:00.000000000',
           '2020-02-01T08:00:00.000000000', '2020-02-01T08:00:00.000000000',
           '2020-02-01T08:00:00.000000000', '2020-02-01T08:00:00.000000000',
           '2020-02-01T08:00:00.000000000', '2020-02-01T09:00:00.000000000',
           '2020-02-01T09:00:00.000000000', '2020-02-01T10:00:00.000000000',
           '2020-02-01T10:00:00.000000000', '2020-02-01T10:00:00.000000000',
           '2020-02-01T10:00:00.000000000', '2020-02-01T10:00:00.000000000',
           '2020-02-01T10:00:00.000000000', '2020-02-01T10:00:00.000000000',
           '2020-02-01T10:00:00.000000000', '2020-02-01T10:00:00.000000000',
           '2020-02-01T11:00:00.000000000', '2020-02-01T11:00:00.000000000',
           '2020-02-01T11:00:00.000000000', '2020-02-01T11:00:00.000000000',
           '2020-02-01T11:00:00.000000000', '2020-02-01T11:00:00.000000000',
           '2020-02-01T12:00:00.000000000', '2020-02-01T12:00:00.000000000',
           '2020-02-01T12:00:00.000000000', '2020-02-01T12:00:00.000000000'],
          dtype='datetime64[ns]')

  • end_basetime
    (tracks)
    datetime64[ns]
    2020-02-01T19:00:00 ... 2020-02-...

    long_name :

    End Epoch time of each track

    array(['2020-02-01T19:00:00.000000000', '2020-02-01T16:00:00.000000000',
           '2020-02-04T22:00:00.000000000', '2020-02-01T21:00:00.000000000',
           '2020-02-02T03:00:00.000000000', '2020-02-02T09:00:00.000000000',
           '2020-02-02T15:00:00.000000000', '2020-02-01T13:00:00.000000000',
           '2020-02-03T09:00:00.000000000', '2020-02-03T04:00:00.000000000',
           '2020-02-02T14:00:00.000000000', '2020-02-03T13:00:00.000000000',
           '2020-02-01T13:00:00.000000000', '2020-02-02T12:00:00.000000000',
           '2020-02-01T16:00:00.000000000', '2020-02-01T12:00:00.000000000',
           '2020-02-02T15:00:00.000000000', '2020-02-02T21:00:00.000000000',
           '2020-02-02T15:00:00.000000000', '2020-02-01T12:00:00.000000000',
           '2020-02-01T12:00:00.000000000', '2020-02-01T15:00:00.000000000',
           '2020-02-01T17:00:00.000000000', '2020-02-03T03:00:00.000000000',
           '2020-02-01T17:00:00.000000000', '2020-02-01T14:00:00.000000000',
           '2020-02-01T13:00:00.000000000', '2020-02-03T01:00:00.000000000',
           '2020-02-02T03:00:00.000000000', '2020-02-02T21:00:00.000000000',
           '2020-02-01T12:00:00.000000000', '2020-02-01T18:00:00.000000000',
           '2020-02-01T13:00:00.000000000', '2020-02-02T01:00:00.000000000',
           '2020-02-01T23:00:00.000000000', '2020-02-01T22:00:00.000000000',
           '2020-02-01T14:00:00.000000000', '2020-02-01T12:00:00.000000000',
           '2020-02-02T06:00:00.000000000', '2020-02-01T13:00:00.000000000',
    ...
           '2020-02-01T13:00:00.000000000', '2020-02-03T13:00:00.000000000',
           '2020-02-01T20:00:00.000000000', '2020-02-02T13:00:00.000000000',
           '2020-02-01T23:00:00.000000000', '2020-02-01T15:00:00.000000000',
           '2020-02-03T13:00:00.000000000', '2020-02-01T12:00:00.000000000',
           '2020-02-02T04:00:00.000000000', '2020-02-01T21:00:00.000000000',
           '2020-02-01T17:00:00.000000000', '2020-02-01T20:00:00.000000000',
           '2020-02-01T17:00:00.000000000', '2020-02-01T16:00:00.000000000',
           '2020-02-03T23:00:00.000000000', '2020-02-01T17:00:00.000000000',
           '2020-02-02T07:00:00.000000000', '2020-02-01T16:00:00.000000000',
           '2020-02-01T21:00:00.000000000', '2020-02-01T17:00:00.000000000',
           '2020-02-01T14:00:00.000000000', '2020-02-02T17:00:00.000000000',
           '2020-02-01T15:00:00.000000000', '2020-02-02T01:00:00.000000000',
           '2020-02-01T21:00:00.000000000', '2020-02-01T18:00:00.000000000',
           '2020-02-01T22:00:00.000000000', '2020-02-01T21:00:00.000000000',
           '2020-02-03T16:00:00.000000000', '2020-02-01T21:00:00.000000000',
           '2020-02-02T02:00:00.000000000', '2020-02-02T12:00:00.000000000',
           '2020-02-02T03:00:00.000000000', '2020-02-01T21:00:00.000000000',
           '2020-02-01T21:00:00.000000000', '2020-02-02T16:00:00.000000000',
           '2020-02-01T22:00:00.000000000', '2020-02-02T10:00:00.000000000'],
          dtype='datetime64[ns]')

  • start_split_cloudnumber
    (tracks)
    float64
    ...

    long_name :

    Cloud number where this track splits from

    units :

    unitless

    [100 values with dtype=float64]

  • end_merge_cloudnumber
    (tracks)
    float64
    ...

    long_name :

    Cloud number where this track merges with

    units :

    unitless

    [100 values with dtype=float64]

  • pbc_flag
    (tracks)
    int32
    ...

    long_name :

    Flag indicating if position adjustment was applied due to PBC

    units :

    unitless

    comments :

    1: adjustment applied, 0: no adjustment

    [100 values with dtype=int32]

  • base_time
    (tracks, times)
    datetime64[ns]
    2020-01-27T21:00:00 ... NaT

    long_name :

    Epoch time of a feature

    array([['2020-01-27T21:00:00.000000000', '2020-01-27T22:00:00.000000000',
            '2020-01-27T23:00:00.000000000', ...,
                                      'NaT',                           'NaT',
                                      'NaT'],
           ['2020-01-29T20:00:00.000000000', '2020-01-29T21:00:00.000000000',
            '2020-01-29T22:00:00.000000000', ...,
                                      'NaT',                           'NaT',
                                      'NaT'],
           ['2020-01-29T22:00:00.000000000', '2020-01-29T23:00:00.000000000',
            '2020-01-30T00:00:00.000000000', ...,
                                      'NaT',                           'NaT',
                                      'NaT'],
           ...,
           ['2020-02-01T12:00:00.000000000', '2020-02-01T13:00:00.000000000',
            '2020-02-01T14:00:00.000000000', ...,
                                      'NaT',                           'NaT',
                                      'NaT'],
           ['2020-02-01T12:00:00.000000000', '2020-02-01T13:00:00.000000000',
            '2020-02-01T14:00:00.000000000', ...,
                                      'NaT',                           'NaT',
                                      'NaT'],
           ['2020-02-01T12:00:00.000000000', '2020-02-01T13:00:00.000000000',
            '2020-02-01T14:00:00.000000000', ...,
                                      'NaT',                           'NaT',
                                      'NaT']], dtype='datetime64[ns]')

  • meanlat
    (tracks, times)
    float32
    ...

    long_name :

    Mean latitude of a feature

    units :

    degrees_north

    [40000 values with dtype=float32]

  • meanlon
    (tracks, times)
    float32
    ...

    long_name :

    Mean longitude of a feature

    units :

    degrees_east

    [40000 values with dtype=float32]

  • area
    (tracks, times)
    float32
    ...

    long_name :

    Area of a feature

    units :

    km^2

    [40000 values with dtype=float32]

  • cloudnumber
    (tracks, times)
    float64
    ...

    long_name :

    Corresponding cloud number in cloudid file

    units :

    unitless

    [40000 values with dtype=float64]

  • track_status
    (tracks, times)
    float64
    ...

    long_name :

    Flag indicating the status of a track

    units :

    unitless

    comments :

    0: Track stops; 1: Simple track continuation; 2: This is the bigger cloud in simple merger; 3: This is the bigger cloud from a simple split that stops at this time; 4: This is the bigger cloud from a split and this cloud continues to the next time; 5: This is the bigger cloud from a split that subsequently is the big cloud in a merger; 13: This cloud splits at the next time step; 15: This cloud is the bigger cloud in a merge that then splits at the next time step; 16: This is the bigger cloud in a split that then splits at the next time step; 18: Merge-split at same time (big merge, splitter, and big split); 21: This is the smaller cloud in a simple merger; 24: This is the bigger cloud of a split that is then the small cloud in a merger; 31: This is the smaller cloud in a simple split that stops; 32: This is a small split that continues onto the next time step; 33: This is a small split that then is the bigger cloud in a merger; 34: This is the small cloud in a merger that then splits at the next time step; 37: Merge-split at same time (small merge, splitter, big split); 44: This is the smaller cloud in a split that is smaller cloud in a merger at the next time step; 46: Merge-split at same time (big merge, splitter, small split); 52: This is the smaller cloud in a split that is smaller cloud in a merger at the next time step; 65: Merge-split at same time (smaller merge, splitter, small split)

    [40000 values with dtype=float64]

  • track_interruptions
    (tracks, times)
    float64
    ...

    long_name :

    0 = full track available, good data. 1 = track starts at first file, track cut short by data availability. 2 = track ends at last file, track cut short by data availability

    units :

    unitless

    [40000 values with dtype=float64]

  • core_area
    (tracks, times)
    float32
    ...

    long_name :

    Area of cold core

    units :

    km^2

    [40000 values with dtype=float32]

  • cold_area
    (tracks, times)
    float32
    ...

    long_name :

    Area of cold anvil

    units :

    km^2

    [40000 values with dtype=float32]

  • corecold_mintb
    (tracks, times)
    float32
    ...

    long_name :

    Minimum Tb in cold core + cold anvil area

    units :

    K

    [40000 values with dtype=float32]

  • corecold_meantb
    (tracks, times)
    float32
    ...

    long_name :

    Mean Tb in cold core + cold anvil area

    units :

    K

    [40000 values with dtype=float32]

  • core_meantb
    (tracks, times)
    float32
    ...

    long_name :

    Mean Tb in cold core area

    units :

    K

    [40000 values with dtype=float32]

  • lat_mintb
    (tracks, times)
    float32
    ...

    long_name :

    Latitude with min Tb

    units :

    degree

    [40000 values with dtype=float32]

  • lon_mintb
    (tracks, times)
    float32
    ...

    long_name :

    Longitude with min Tb

    units :

    degree

    [40000 values with dtype=float32]

  • mcs_duration
    (tracks)
    int64
    ...

    long_name :

    Duration of MCS stage

    units :

    unitless

    comments :

    Multiply by time_resolution_hour to convert to physical units

    [100 values with dtype=int64]

  • mcs_status
    (tracks, times)
    float32
    ...

    long_name :

    Flag indicating the status of MCS based on Tb. 1 = Yes, 0 = No

    units :

    unitless

    [40000 values with dtype=float32]

  • ccs_area
    (tracks, times)
    float32
    4.56e+04 5.3e+04 ... nan nan

    long_name :

    Area of cold cloud shield

    units :

    km^2

    array([[ 45600.,  53000.,  50300., ...,     nan,     nan,     nan],
           [304100., 239700., 356000., ...,     nan,     nan,     nan],
           [ 29900.,  16300.,  21500., ...,     nan,     nan,     nan],
           ...,
           [ 35800., 115100., 141300., ...,     nan,     nan,     nan],
           [ 14400.,  14700.,  18100., ...,     nan,     nan,     nan],
           [  5400.,  10500.,  14400., ...,     nan,     nan,     nan]],
          dtype=float32)

  • merge_cloudnumber
    (tracks, times, mergers)
    float64
    ...

    long_name :

    Cloud numbers that merge into MCS

    units :

    unitless

    [2000000 values with dtype=float64]

  • split_cloudnumber
    (tracks, times, mergers)
    float64
    ...

    long_name :

    Cloud numbers that split from MCS

    units :

    unitless

    [2000000 values with dtype=float64]

  • merge_ccs_area
    (tracks, times, mergers)
    float32
    ...

    long_name :

    Cold cloud shield area that merge into MCS

    units :

    km^2

    [2000000 values with dtype=float32]

  • split_ccs_area
    (tracks, times, mergers)
    float32
    ...

    long_name :

    Cold cloud shield area that split from MCS

    units :

    km^2

    [2000000 values with dtype=float32]

  • pf_npf
    (tracks, times)
    float32
    ...

    long_name :

    Number of PF in the cloud

    units :

    unitless

    [40000 values with dtype=float32]

  • pf_lon
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Mean longitude of PF

    units :

    degrees

    [120000 values with dtype=float32]

  • pf_lat
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Mean latitude of PF

    units :

    degrees

    [120000 values with dtype=float32]

  • pf_area
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Area of PF

    units :

    km^2

    [120000 values with dtype=float32]

  • pf_rainrate
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Mean rain rate of PF

    units :

    mm/h

    [120000 values with dtype=float32]

  • pf_skewness
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Rain rate skewness of PF

    units :

    unitless

    [120000 values with dtype=float32]

  • pf_majoraxis
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Major axis length of PF

    units :

    km

    [120000 values with dtype=float32]

  • pf_minoraxis
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Minor axis length of PF

    units :

    km

    [120000 values with dtype=float32]

  • pf_aspectratio
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Aspect ratio (major axis / minor axis) of PF

    units :

    unitless

    [120000 values with dtype=float32]

  • pf_orientation
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Orientation of major axis of PF

    units :

    degrees

    [120000 values with dtype=float32]

  • pf_perimeter
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Perimeter of PF

    units :

    km

    [120000 values with dtype=float32]

  • pf_eccentricity
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Eccentricity of PF

    units :

    unitless

    [120000 values with dtype=float32]

  • pf_lon_centroid
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Centroid longitude of PF

    units :

    degrees

    [120000 values with dtype=float32]

  • pf_lat_centroid
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Centroid latitude of PF

    units :

    degrees

    [120000 values with dtype=float32]

  • pf_lon_weightedcentroid
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Weighted centroid longitude of PF

    units :

    degrees

    [120000 values with dtype=float32]

  • pf_lat_weightedcentroid
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Weighted centroid latitude of PF

    units :

    degrees

    [120000 values with dtype=float32]

  • pf_lon_maxrainrate
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Longitude with max rain rate

    units :

    degree

    [120000 values with dtype=float32]

  • pf_lat_maxrainrate
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Latitude with max rain rate

    units :

    degree

    [120000 values with dtype=float32]

  • pf_maxrainrate
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Max rain rate of PF

    units :

    mm/h

    [120000 values with dtype=float32]

  • pf_accumrain
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Accumulate precipitation of PF

    units :

    mm/h

    [120000 values with dtype=float32]

  • pf_accumrainheavy
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Accumulated heavy precipitation of PF

    units :

    mm/h

    heavy_rainrate_threshold :

    10.0

    [120000 values with dtype=float32]

  • pf_landfrac
    (tracks, times)
    float32
    ...

    long_name :

    Fraction of PF over land

    units :

    fraction

    [40000 values with dtype=float32]

  • total_rain
    (tracks, times)
    float32
    ...

    long_name :

    Total precipitation under cold cloud shield (all rainfall included)

    units :

    mm/h

    [40000 values with dtype=float32]

  • total_heavyrain
    (tracks, times)
    float32
    ...

    long_name :

    Total heavy precipitation under cold cloud shield

    units :

    mm/h

    heavy_rainrate_threshold :

    10.0

    [40000 values with dtype=float32]

  • rainrate_heavyrain
    (tracks, times)
    float32
    ...

    long_name :

    Mean heavy rain rate under cold cloud shield

    units :

    mm/h

    heavy_rainrate_threshold :

    10.0

    [40000 values with dtype=float32]

  • pf_lifetime
    (tracks)
    float32
    ...

    long_name :

    MCS lifetime when a significant PF is present

    units :

    hour

    [100 values with dtype=float32]

  • pf_mcsstatus
    (tracks, times)
    float64
    ...

    long_name :

    Flag indicating the status of MCS based on PF. 1 = Yes, 0 = No

    units :

    unitless

    [40000 values with dtype=float64]

  • movement_distance
    (tracks, times)
    float32
    ...

    long_name :

    Movement distance along angle movement_theta

    units :

    km

    lag :

    1

    comments :

    This is the total movement along the angle theta between lag estimates.

    [40000 values with dtype=float32]

  • movement_speed
    (tracks, times)
    float32
    ...

    long_name :

    Movement speed along angle movement_theta

    units :

    m/s

    [40000 values with dtype=float32]

  • movement_theta
    (tracks, times)
    float32
    ...

    long_name :

    Movement direction

    units :

    degrees

    comments :

    0:northward, 90:eastward, 180:southward, 270:westward.

    [40000 values with dtype=float32]

  • movement_distance_x
    (tracks, times)
    float32
    ...

    long_name :

    East-West component of movement distance

    units :

    km

    [40000 values with dtype=float32]

  • movement_distance_y
    (tracks, times)
    float32
    ...

    long_name :

    North-South component of movement distance

    units :

    km

    [40000 values with dtype=float32]

• Indexes: (4)
  • tracks
    PandasIndex

    PandasIndex(Index([20832, 21069, 21085, 21099, 21166, 21199, 21205, 21210, 21217, 21221,
           21231, 21248, 21260, 21271, 21290, 21294, 21300, 21304, 21319, 21328,
           21329, 21336, 21337, 21341, 21344, 21345, 21347, 21348, 21352, 21356,
           21358, 21360, 21363, 21364, 21365, 21367, 21370, 21372, 21373, 21374,
           21375, 21376, 21380, 21381, 21382, 21383, 21386, 21393, 21394, 21395,
           21396, 21398, 21399, 21400, 21401, 21402, 21403, 21404, 21405, 21406,
           21407, 21408, 21409, 21410, 21411, 21412, 21413, 21415, 21416, 21417,
           21418, 21419, 21420, 21421, 21422, 21423, 21424, 21425, 21426, 21427,
           21428, 21429, 21430, 21431, 21432, 21433, 21434, 21435, 21436, 21437,
           21438, 21439, 21440, 21441, 21442, 21443, 21444, 21445, 21446, 21447],
          dtype='int64', name='tracks'))

  • times
    PandasIndex

    PandasIndex(Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
           ...
           390, 391, 392, 393, 394, 395, 396, 397, 398, 399],
          dtype='int64', name='times', length=400))

  • mergers
    PandasIndex

    PandasIndex(Index([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16, 17,
           18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35,
           36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49],
          dtype='int64', name='mergers'))

  • nmaxpf
    PandasIndex

    PandasIndex(Index([0, 1, 2], dtype='int64', name='nmaxpf'))

• Attributes: (21)

  Title :

  Statistics of each MCS track

  Institution :

  Pacific Northwest National Laboratory

  Contact :

  Zhe Feng, zhe.feng@pnnl.gov

  Created_on :

  Sun Apr 13 21:03:31 2025

  startdate :

  20190901.0000

  enddate :

  20200901.0000

  timegap_hour :

  3.1

  time_resolution_hour :

  1.0

  pixel_radius_km :

  10.0

  MCS_duration_hr :

  4

  MCS_area_km^2 :

  40000

  nmaxpf :

  3

  PF_rainrate_thresh :

  2.0

  heavy_rainrate_thresh :

  10.0

  landfrac_thresh :

  [ 0 90]

  MCS_PF_majoraxis_thresh :

  0

  MCS_PF_duration_thresh :

  4

  PF_PF_min_majoraxis_thresh :

  20

  mcs_min_rainvol_thresh :

  20000

  mcs_volrain_duration_thresh :

  1.0

  max_speed_thresh :

  50

# Or access tracks based on their location:
# N.B. force a load of meanlat.
dstracks.meanlat.load()
# This suppresses a warning about chunk sizes.
with dask.config.set(**{'array.slicing.split_large_chunks': False}):
    dstracks_tropical = dstracks.isel(
        tracks=((dstracks.meanlat.values > -20) & (dstracks.meanlat.values < 20)).any(axis=1)
    )
dstracks_tropical

<xarray.Dataset> Size: 21GB
Dimensions:                  (tracks: 34165, times: 400, mergers: 50, nmaxpf: 3)
Coordinates:
  * tracks                   (tracks) int64 273kB 1 4 5 6 ... 52110 52111 52112
  * times                    (times) int64 3kB 0 1 2 3 4 ... 395 396 397 398 399
  * mergers                  (mergers) int64 400B 0 1 2 3 4 5 ... 45 46 47 48 49
  * nmaxpf                   (nmaxpf) int64 24B 0 1 2
Data variables: (12/61)
    track_duration           (tracks) int32 137kB ...
    start_status             (tracks) float64 273kB ...
    end_status               (tracks) float64 273kB ...
    start_basetime           (tracks) datetime64[ns] 273kB 2019-09-01 ... 202...
    end_basetime             (tracks) datetime64[ns] 273kB 2019-09-13T11:00:0...
    start_split_cloudnumber  (tracks) float64 273kB ...
    ...                       ...
    pf_mcsstatus             (tracks, times) float64 109MB ...
    movement_distance        (tracks, times) float32 55MB ...
    movement_speed           (tracks, times) float32 55MB ...
    movement_theta           (tracks, times) float32 55MB ...
    movement_distance_x      (tracks, times) float32 55MB ...
    movement_distance_y      (tracks, times) float32 55MB ...
Attributes: (12/21)
    Title:                        Statistics of each MCS track
    Institution:                  Pacific Northwest National Laboratory
    Contact:                      Zhe Feng, zhe.feng@pnnl.gov
    Created_on:                   Sun Apr 13 21:03:31 2025
    startdate:                    20190901.0000
    enddate:                      20200901.0000
    ...                           ...
    MCS_PF_majoraxis_thresh:      0
    MCS_PF_duration_thresh:       4
    PF_PF_min_majoraxis_thresh:   20
    mcs_min_rainvol_thresh:       20000
    mcs_volrain_duration_thresh:  1.0
    max_speed_thresh:             50

xarray.Dataset

• Dimensions:
  • tracks: 34165
  • times: 400
  • mergers: 50
  • nmaxpf: 3
• Coordinates: (4)
  • tracks
    (tracks)
    int64
    1 4 5 6 ... 52108 52110 52111 52112

    array([    1,     4,     5, ..., 52110, 52111, 52112])

  • times
    (times)
    int64
    0 1 2 3 4 5 ... 395 396 397 398 399

    array([  0,   1,   2, ..., 397, 398, 399])

  • mergers
    (mergers)
    int64
    0 1 2 3 4 5 6 ... 44 45 46 47 48 49

    array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16, 17,
           18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35,
           36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49])

  • nmaxpf
    (nmaxpf)
    int64
    0 1 2

    array([0, 1, 2])

• Data variables: (61)
  • track_duration
    (tracks)
    int32
    ...

    long_name :

    Duration of each track

    units :

    unitless

    comments :

    Multiply by time_resolution_hour to convert to physical units

    [34165 values with dtype=int32]

  • start_status
    (tracks)
    float64
    ...

    long_name :

    Flag indicating how the track starts

    units :

    unitless

    comments :

    Refer to track_status attributes on flag value meaning

    [34165 values with dtype=float64]

  • end_status
    (tracks)
    float64
    ...

    long_name :

    Flag indicating how the track ends

    units :

    unitless

    comments :

    Refer to track_status attributes on flag value meaning

    [34165 values with dtype=float64]

  • start_basetime
    (tracks)
    datetime64[ns]
    2019-09-01 ... 2020-08-31T20:00:00

    long_name :

    Start Epoch time of each track

    array(['2019-09-01T00:00:00.000000000', '2019-09-01T00:00:00.000000000',
           '2019-09-01T00:00:00.000000000', ...,
           '2020-08-31T19:00:00.000000000', '2020-08-31T20:00:00.000000000',
           '2020-08-31T20:00:00.000000000'], dtype='datetime64[ns]')

  • end_basetime
    (tracks)
    datetime64[ns]
    2019-09-13T11:00:00 ... 2020-09-01

    long_name :

    End Epoch time of each track

    array(['2019-09-13T11:00:00.000000000', '2019-09-11T22:00:00.000000000',
           '2019-09-02T08:00:00.000000000', ...,
           '2020-09-01T00:00:00.000000000', '2020-09-01T00:00:00.000000000',
           '2020-09-01T00:00:00.000000000'], dtype='datetime64[ns]')

  • start_split_cloudnumber
    (tracks)
    float64
    ...

    long_name :

    Cloud number where this track splits from

    units :

    unitless

    [34165 values with dtype=float64]

  • end_merge_cloudnumber
    (tracks)
    float64
    ...

    long_name :

    Cloud number where this track merges with

    units :

    unitless

    [34165 values with dtype=float64]

  • pbc_flag
    (tracks)
    int32
    ...

    long_name :

    Flag indicating if position adjustment was applied due to PBC

    units :

    unitless

    comments :

    1: adjustment applied, 0: no adjustment

    [34165 values with dtype=int32]

  • base_time
    (tracks, times)
    datetime64[ns]
    2019-09-01 ... NaT

    long_name :

    Epoch time of a feature

    array([['2019-09-01T00:00:00.000000000', '2019-09-01T01:00:00.000000000',
            '2019-09-01T02:00:00.000000000', ...,
                                      'NaT',                           'NaT',
                                      'NaT'],
           ['2019-09-01T00:00:00.000000000', '2019-09-01T01:00:00.000000000',
            '2019-09-01T02:00:00.000000000', ...,
                                      'NaT',                           'NaT',
                                      'NaT'],
           ['2019-09-01T00:00:00.000000000', '2019-09-01T01:00:00.000000000',
            '2019-09-01T02:00:00.000000000', ...,
                                      'NaT',                           'NaT',
                                      'NaT'],
           ...,
           ['2020-08-31T19:00:00.000000000', '2020-08-31T20:00:00.000000000',
            '2020-08-31T21:00:00.000000000', ...,
                                      'NaT',                           'NaT',
                                      'NaT'],
           ['2020-08-31T20:00:00.000000000', '2020-08-31T21:00:00.000000000',
            '2020-08-31T22:00:00.000000000', ...,
                                      'NaT',                           'NaT',
                                      'NaT'],
           ['2020-08-31T20:00:00.000000000', '2020-08-31T21:00:00.000000000',
            '2020-08-31T22:00:00.000000000', ...,
                                      'NaT',                           'NaT',
                                      'NaT']], dtype='datetime64[ns]')

  • meanlat
    (tracks, times)
    float32
    12.04 12.33 12.41 ... nan nan nan

    long_name :

    Mean latitude of a feature

    units :

    degrees_north

    array([[12.036651  , 12.329901  , 12.412557  , ...,         nan,
                    nan,         nan],
           [12.845157  , 12.793903  , 12.647314  , ...,         nan,
                    nan,         nan],
           [ 8.274477  ,  8.3808975 ,  8.385168  , ...,         nan,
                    nan,         nan],
           ...,
           [10.56282   , 10.768919  , 10.830226  , ...,         nan,
                    nan,         nan],
           [-0.9659906 , -1.0222977 , -0.81335825, ...,         nan,
                    nan,         nan],
           [ 9.615461  ,  9.556007  ,  9.581835  , ...,         nan,
                    nan,         nan]], dtype=float32)

  • meanlon
    (tracks, times)
    float32
    ...

    long_name :

    Mean longitude of a feature

    units :

    degrees_east

    [13666000 values with dtype=float32]

  • area
    (tracks, times)
    float32
    ...

    long_name :

    Area of a feature

    units :

    km^2

    [13666000 values with dtype=float32]

  • cloudnumber
    (tracks, times)
    float64
    ...

    long_name :

    Corresponding cloud number in cloudid file

    units :

    unitless

    [13666000 values with dtype=float64]

  • track_status
    (tracks, times)
    float64
    ...

    long_name :

    Flag indicating the status of a track

    units :

    unitless

    comments :

    0: Track stops; 1: Simple track continuation; 2: This is the bigger cloud in simple merger; 3: This is the bigger cloud from a simple split that stops at this time; 4: This is the bigger cloud from a split and this cloud continues to the next time; 5: This is the bigger cloud from a split that subsequently is the big cloud in a merger; 13: This cloud splits at the next time step; 15: This cloud is the bigger cloud in a merge that then splits at the next time step; 16: This is the bigger cloud in a split that then splits at the next time step; 18: Merge-split at same time (big merge, splitter, and big split); 21: This is the smaller cloud in a simple merger; 24: This is the bigger cloud of a split that is then the small cloud in a merger; 31: This is the smaller cloud in a simple split that stops; 32: This is a small split that continues onto the next time step; 33: This is a small split that then is the bigger cloud in a merger; 34: This is the small cloud in a merger that then splits at the next time step; 37: Merge-split at same time (small merge, splitter, big split); 44: This is the smaller cloud in a split that is smaller cloud in a merger at the next time step; 46: Merge-split at same time (big merge, splitter, small split); 52: This is the smaller cloud in a split that is smaller cloud in a merger at the next time step; 65: Merge-split at same time (smaller merge, splitter, small split)

    [13666000 values with dtype=float64]

  • track_interruptions
    (tracks, times)
    float64
    ...

    long_name :

    0 = full track available, good data. 1 = track starts at first file, track cut short by data availability. 2 = track ends at last file, track cut short by data availability

    units :

    unitless

    [13666000 values with dtype=float64]

  • core_area
    (tracks, times)
    float32
    ...

    long_name :

    Area of cold core

    units :

    km^2

    [13666000 values with dtype=float32]

  • cold_area
    (tracks, times)
    float32
    ...

    long_name :

    Area of cold anvil

    units :

    km^2

    [13666000 values with dtype=float32]

  • corecold_mintb
    (tracks, times)
    float32
    ...

    long_name :

    Minimum Tb in cold core + cold anvil area

    units :

    K

    [13666000 values with dtype=float32]

  • corecold_meantb
    (tracks, times)
    float32
    ...

    long_name :

    Mean Tb in cold core + cold anvil area

    units :

    K

    [13666000 values with dtype=float32]

  • core_meantb
    (tracks, times)
    float32
    ...

    long_name :

    Mean Tb in cold core area

    units :

    K

    [13666000 values with dtype=float32]

  • lat_mintb
    (tracks, times)
    float32
    ...

    long_name :

    Latitude with min Tb

    units :

    degree

    [13666000 values with dtype=float32]

  • lon_mintb
    (tracks, times)
    float32
    ...

    long_name :

    Longitude with min Tb

    units :

    degree

    [13666000 values with dtype=float32]

  • mcs_duration
    (tracks)
    int64
    ...

    long_name :

    Duration of MCS stage

    units :

    unitless

    comments :

    Multiply by time_resolution_hour to convert to physical units

    [34165 values with dtype=int64]

  • mcs_status
    (tracks, times)
    float32
    ...

    long_name :

    Flag indicating the status of MCS based on Tb. 1 = Yes, 0 = No

    units :

    unitless

    [13666000 values with dtype=float32]

  • ccs_area
    (tracks, times)
    float32
    7.044e+05 6.673e+05 ... nan nan

    long_name :

    Area of cold cloud shield

    units :

    km^2

    array([[704400., 667300., 744000., ...,     nan,     nan,     nan],
           [567800., 607500., 606800., ...,     nan,     nan,     nan],
           [420900., 412200., 384400., ...,     nan,     nan,     nan],
           ...,
           [  3300.,  25300.,  52500., ...,     nan,     nan,     nan],
           [381300., 402900., 280000., ...,     nan,     nan,     nan],
           [ 53000.,  50400.,  54300., ...,     nan,     nan,     nan]],
          dtype=float32)

  • merge_cloudnumber
    (tracks, times, mergers)
    float64
    ...

    long_name :

    Cloud numbers that merge into MCS

    units :

    unitless

    [683300000 values with dtype=float64]

  • split_cloudnumber
    (tracks, times, mergers)
    float64
    ...

    long_name :

    Cloud numbers that split from MCS

    units :

    unitless

    [683300000 values with dtype=float64]

  • merge_ccs_area
    (tracks, times, mergers)
    float32
    ...

    long_name :

    Cold cloud shield area that merge into MCS

    units :

    km^2

    [683300000 values with dtype=float32]

  • split_ccs_area
    (tracks, times, mergers)
    float32
    ...

    long_name :

    Cold cloud shield area that split from MCS

    units :

    km^2

    [683300000 values with dtype=float32]

  • pf_npf
    (tracks, times)
    float32
    ...

    long_name :

    Number of PF in the cloud

    units :

    unitless

    [13666000 values with dtype=float32]

  • pf_lon
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Mean longitude of PF

    units :

    degrees

    [40998000 values with dtype=float32]

  • pf_lat
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Mean latitude of PF

    units :

    degrees

    [40998000 values with dtype=float32]

  • pf_area
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Area of PF

    units :

    km^2

    [40998000 values with dtype=float32]

  • pf_rainrate
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Mean rain rate of PF

    units :

    mm/h

    [40998000 values with dtype=float32]

  • pf_skewness
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Rain rate skewness of PF

    units :

    unitless

    [40998000 values with dtype=float32]

  • pf_majoraxis
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Major axis length of PF

    units :

    km

    [40998000 values with dtype=float32]

  • pf_minoraxis
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Minor axis length of PF

    units :

    km

    [40998000 values with dtype=float32]

  • pf_aspectratio
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Aspect ratio (major axis / minor axis) of PF

    units :

    unitless

    [40998000 values with dtype=float32]

  • pf_orientation
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Orientation of major axis of PF

    units :

    degrees

    [40998000 values with dtype=float32]

  • pf_perimeter
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Perimeter of PF

    units :

    km

    [40998000 values with dtype=float32]

  • pf_eccentricity
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Eccentricity of PF

    units :

    unitless

    [40998000 values with dtype=float32]

  • pf_lon_centroid
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Centroid longitude of PF

    units :

    degrees

    [40998000 values with dtype=float32]

  • pf_lat_centroid
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Centroid latitude of PF

    units :

    degrees

    [40998000 values with dtype=float32]

  • pf_lon_weightedcentroid
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Weighted centroid longitude of PF

    units :

    degrees

    [40998000 values with dtype=float32]

  • pf_lat_weightedcentroid
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Weighted centroid latitude of PF

    units :

    degrees

    [40998000 values with dtype=float32]

  • pf_lon_maxrainrate
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Longitude with max rain rate

    units :

    degree

    [40998000 values with dtype=float32]

  • pf_lat_maxrainrate
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Latitude with max rain rate

    units :

    degree

    [40998000 values with dtype=float32]

  • pf_maxrainrate
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Max rain rate of PF

    units :

    mm/h

    [40998000 values with dtype=float32]

  • pf_accumrain
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Accumulate precipitation of PF

    units :

    mm/h

    [40998000 values with dtype=float32]

  • pf_accumrainheavy
    (tracks, times, nmaxpf)
    float32
    ...

    long_name :

    Accumulated heavy precipitation of PF

    units :

    mm/h

    heavy_rainrate_threshold :

    10.0

    [40998000 values with dtype=float32]

  • pf_landfrac
    (tracks, times)
    float32
    ...

    long_name :

    Fraction of PF over land

    units :

    fraction

    [13666000 values with dtype=float32]

  • total_rain
    (tracks, times)
    float32
    ...

    long_name :

    Total precipitation under cold cloud shield (all rainfall included)

    units :

    mm/h

    [13666000 values with dtype=float32]

  • total_heavyrain
    (tracks, times)
    float32
    ...

    long_name :

    Total heavy precipitation under cold cloud shield

    units :

    mm/h

    heavy_rainrate_threshold :

    10.0

    [13666000 values with dtype=float32]

  • rainrate_heavyrain
    (tracks, times)
    float32
    ...

    long_name :

    Mean heavy rain rate under cold cloud shield

    units :

    mm/h

    heavy_rainrate_threshold :

    10.0

    [13666000 values with dtype=float32]

  • pf_lifetime
    (tracks)
    float32
    ...

    long_name :

    MCS lifetime when a significant PF is present

    units :

    hour

    [34165 values with dtype=float32]

  • pf_mcsstatus
    (tracks, times)
    float64
    ...

    long_name :

    Flag indicating the status of MCS based on PF. 1 = Yes, 0 = No

    units :

    unitless

    [13666000 values with dtype=float64]

  • movement_distance
    (tracks, times)
    float32
    ...

    long_name :

    Movement distance along angle movement_theta

    units :

    km

    lag :

    1

    comments :

    This is the total movement along the angle theta between lag estimates.

    [13666000 values with dtype=float32]

  • movement_speed
    (tracks, times)
    float32
    ...

    long_name :

    Movement speed along angle movement_theta

    units :

    m/s

    [13666000 values with dtype=float32]

  • movement_theta
    (tracks, times)
    float32
    ...

    long_name :

    Movement direction

    units :

    degrees

    comments :

    0:northward, 90:eastward, 180:southward, 270:westward.

    [13666000 values with dtype=float32]

  • movement_distance_x
    (tracks, times)
    float32
    ...

    long_name :

    East-West component of movement distance

    units :

    km

    [13666000 values with dtype=float32]

  • movement_distance_y
    (tracks, times)
    float32
    ...

    long_name :

    North-South component of movement distance

    units :

    km

    [13666000 values with dtype=float32]

• Indexes: (4)
  • tracks
    PandasIndex

    PandasIndex(Index([    1,     4,     5,     6,     9,    10,    11,    14,    16,    17,
           ...
           52095, 52098, 52101, 52102, 52106, 52107, 52108, 52110, 52111, 52112],
          dtype='int64', name='tracks', length=34165))

  • times
    PandasIndex

    PandasIndex(Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
           ...
           390, 391, 392, 393, 394, 395, 396, 397, 398, 399],
          dtype='int64', name='times', length=400))

  • mergers
    PandasIndex

    PandasIndex(Index([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16, 17,
           18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35,
           36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49],
          dtype='int64', name='mergers'))

  • nmaxpf
    PandasIndex

    PandasIndex(Index([0, 1, 2], dtype='int64', name='nmaxpf'))

• Attributes: (21)

  Title :

  Statistics of each MCS track

  Institution :

  Pacific Northwest National Laboratory

  Contact :

  Zhe Feng, zhe.feng@pnnl.gov

  Created_on :

  Sun Apr 13 21:03:31 2025

  startdate :

  20190901.0000

  enddate :

  20200901.0000

  timegap_hour :

  3.1

  time_resolution_hour :

  1.0

  pixel_radius_km :

  10.0

  MCS_duration_hr :

  4

  MCS_area_km^2 :

  40000

  nmaxpf :

  3

  PF_rainrate_thresh :

  2.0

  heavy_rainrate_thresh :

  10.0

  landfrac_thresh :

  [ 0 90]

  MCS_PF_majoraxis_thresh :

  0

  MCS_PF_duration_thresh :

  4

  PF_PF_min_majoraxis_thresh :

  20

  mcs_min_rainvol_thresh :

  20000

  mcs_volrain_duration_thresh :

  1.0

  max_speed_thresh :

  50

# Each track can then be looped over using e.g.:
for track_id in dstracks_at_time.tracks.values[:10]:  # Just get first 10.
    track = dstracks_at_time.sel(tracks=track_id)
    print(f"Track ID: {track.tracks.values.item()}, duration: {track.track_duration.values.item()}")

Track ID: 20832, duration: 119
Track ID: 21069, duration: 69
Track ID: 21085, duration: 145
Track ID: 21099, duration: 68
Track ID: 21166, duration: 62
Track ID: 21199, duration: 64
Track ID: 21205, duration: 69
Track ID: 21210, duration: 42
Track ID: 21217, duration: 85
Track ID: 21221, duration: 79

Individual track properties

# Select a track.
track = dstracks.sel(tracks=2001)
track

<xarray.Dataset> Size: 631kB
Dimensions:                  (times: 400, mergers: 50, nmaxpf: 3)
Coordinates:
    tracks                   int64 8B 2001
  * times                    (times) int64 3kB 0 1 2 3 4 ... 395 396 397 398 399
  * mergers                  (mergers) int64 400B 0 1 2 3 4 5 ... 45 46 47 48 49
  * nmaxpf                   (nmaxpf) int64 24B 0 1 2
Data variables: (12/61)
    track_duration           int32 4B ...
    start_status             float64 8B ...
    end_status               float64 8B ...
    start_basetime           datetime64[ns] 8B 2019-09-17T01:00:00
    end_basetime             datetime64[ns] 8B 2019-09-18T03:00:00
    start_split_cloudnumber  float64 8B ...
    ...                       ...
    pf_mcsstatus             (times) float64 3kB ...
    movement_distance        (times) float32 2kB ...
    movement_speed           (times) float32 2kB ...
    movement_theta           (times) float32 2kB ...
    movement_distance_x      (times) float32 2kB ...
    movement_distance_y      (times) float32 2kB ...
Attributes: (12/21)
    Title:                        Statistics of each MCS track
    Institution:                  Pacific Northwest National Laboratory
    Contact:                      Zhe Feng, zhe.feng@pnnl.gov
    Created_on:                   Sun Apr 13 21:03:31 2025
    startdate:                    20190901.0000
    enddate:                      20200901.0000
    ...                           ...
    MCS_PF_majoraxis_thresh:      0
    MCS_PF_duration_thresh:       4
    PF_PF_min_majoraxis_thresh:   20
    mcs_min_rainvol_thresh:       20000
    mcs_volrain_duration_thresh:  1.0
    max_speed_thresh:             50

xarray.Dataset

• Dimensions:
  • times: 400
  • mergers: 50
  • nmaxpf: 3
• Coordinates: (4)
  • tracks
    ()
    int64
    2001

    array(2001)

  • times
    (times)
    int64
    0 1 2 3 4 5 ... 395 396 397 398 399

    array([  0,   1,   2, ..., 397, 398, 399])

  • mergers
    (mergers)
    int64
    0 1 2 3 4 5 6 ... 44 45 46 47 48 49

    array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16, 17,
           18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35,
           36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49])

  • nmaxpf
    (nmaxpf)
    int64
    0 1 2

    array([0, 1, 2])

• Data variables: (61)
  • track_duration
    ()
    int32
    ...

    long_name :

    Duration of each track

    units :

    unitless

    comments :

    Multiply by time_resolution_hour to convert to physical units

    [1 values with dtype=int32]

  • start_status
    ()
    float64
    ...

    long_name :

    Flag indicating how the track starts

    units :

    unitless

    comments :

    Refer to track_status attributes on flag value meaning

    [1 values with dtype=float64]

  • end_status
    ()
    float64
    ...

    long_name :

    Flag indicating how the track ends

    units :

    unitless

    comments :

    Refer to track_status attributes on flag value meaning

    [1 values with dtype=float64]

  • start_basetime
    ()
    datetime64[ns]
    2019-09-17T01:00:00

    long_name :

    Start Epoch time of each track

    array('2019-09-17T01:00:00.000000000', dtype='datetime64[ns]')

  • end_basetime
    ()
    datetime64[ns]
    2019-09-18T03:00:00

    long_name :

    End Epoch time of each track

    array('2019-09-18T03:00:00.000000000', dtype='datetime64[ns]')

  • start_split_cloudnumber
    ()
    float64
    ...

    long_name :

    Cloud number where this track splits from

    units :

    unitless

    [1 values with dtype=float64]

  • end_merge_cloudnumber
    ()
    float64
    ...

    long_name :

    Cloud number where this track merges with

    units :

    unitless

    [1 values with dtype=float64]

  • pbc_flag
    ()
    int32
    ...

    long_name :

    Flag indicating if position adjustment was applied due to PBC

    units :

    unitless

    comments :

    1: adjustment applied, 0: no adjustment

    [1 values with dtype=int32]

  • base_time
    (times)
    datetime64[ns]
    2019-09-17T01:00:00 ... NaT

    long_name :

    Epoch time of a feature

    array(['2019-09-17T01:00:00.000000000', '2019-09-17T02:00:00.000000000',
           '2019-09-17T03:00:00.000000000', '2019-09-17T04:00:00.000000000',
           '2019-09-17T05:00:00.000000000', '2019-09-17T06:00:00.000000000',
           '2019-09-17T07:00:00.000000000', '2019-09-17T08:00:00.000000000',
           '2019-09-17T09:00:00.000000000', '2019-09-17T10:00:00.000000000',
           '2019-09-17T11:00:00.000000000', '2019-09-17T12:00:00.000000000',
           '2019-09-17T13:00:00.000000000', '2019-09-17T14:00:00.000000000',
           '2019-09-17T15:00:00.000000000', '2019-09-17T16:00:00.000000000',
           '2019-09-17T17:00:00.000000000', '2019-09-17T18:00:00.000000000',
           '2019-09-17T19:00:00.000000000', '2019-09-17T20:00:00.000000000',
           '2019-09-17T21:00:00.000000000', '2019-09-17T22:00:00.000000000',
           '2019-09-17T23:00:00.000000000', '2019-09-18T00:00:00.000000000',
           '2019-09-18T01:00:00.000000000', '2019-09-18T02:00:00.000000000',
           '2019-09-18T03:00:00.000000000',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
    ...
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT',
                                     'NaT',                           'NaT'],
          dtype='datetime64[ns]')

  • meanlat
    (times)
    float32
    16.54 16.54 16.64 ... nan nan nan

    long_name :

    Mean latitude of a feature

    units :

    degrees_north

    array([16.536486 , 16.535612 , 16.64209  , 16.080866 , 15.5087595,
           15.653267 , 15.742958 , 15.961675 , 16.138716 , 16.223557 ,
           16.275038 , 16.479338 , 16.64655  , 16.184103 , 16.799896 ,
           15.056975 , 13.404202 , 11.734979 , 13.182419 , 12.09722  ,
           12.7988825, 13.461313 , 12.205271 , 13.543533 , 12.721694 ,
           14.339735 , 12.126923 ,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
    ...
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan,
                  nan,        nan,        nan,        nan,        nan],
          dtype=float32)

  • meanlon
    (times)
    float32
    ...

    long_name :

    Mean longitude of a feature

    units :

    degrees_east

    [400 values with dtype=float32]

  • area
    (times)
    float32
    ...

    long_name :

    Area of a feature

    units :

    km^2

    [400 values with dtype=float32]

  • cloudnumber
    (times)
    float64
    ...

    long_name :

    Corresponding cloud number in cloudid file

    units :

    unitless

    [400 values with dtype=float64]

  • track_status
    (times)
    float64
    ...

    long_name :

    Flag indicating the status of a track

    units :

    unitless

    comments :

    0: Track stops; 1: Simple track continuation; 2: This is the bigger cloud in simple merger; 3: This is the bigger cloud from a simple split that stops at this time; 4: This is the bigger cloud from a split and this cloud continues to the next time; 5: This is the bigger cloud from a split that subsequently is the big cloud in a merger; 13: This cloud splits at the next time step; 15: This cloud is the bigger cloud in a merge that then splits at the next time step; 16: This is the bigger cloud in a split that then splits at the next time step; 18: Merge-split at same time (big merge, splitter, and big split); 21: This is the smaller cloud in a simple merger; 24: This is the bigger cloud of a split that is then the small cloud in a merger; 31: This is the smaller cloud in a simple split that stops; 32: This is a small split that continues onto the next time step; 33: This is a small split that then is the bigger cloud in a merger; 34: This is the small cloud in a merger that then splits at the next time step; 37: Merge-split at same time (small merge, splitter, big split); 44: This is the smaller cloud in a split that is smaller cloud in a merger at the next time step; 46: Merge-split at same time (big merge, splitter, small split); 52: This is the smaller cloud in a split that is smaller cloud in a merger at the next time step; 65: Merge-split at same time (smaller merge, splitter, small split)

    [400 values with dtype=float64]

  • track_interruptions
    (times)
    float64
    ...

    long_name :

    0 = full track available, good data. 1 = track starts at first file, track cut short by data availability. 2 = track ends at last file, track cut short by data availability

    units :

    unitless

    [400 values with dtype=float64]

  • core_area
    (times)
    float32
    ...

    long_name :

    Area of cold core

    units :

    km^2

    [400 values with dtype=float32]

  • cold_area
    (times)
    float32
    ...

    long_name :

    Area of cold anvil

    units :

    km^2

    [400 values with dtype=float32]

  • corecold_mintb
    (times)
    float32
    ...

    long_name :

    Minimum Tb in cold core + cold anvil area

    units :

    K

    [400 values with dtype=float32]

  • corecold_meantb
    (times)
    float32
    ...

    long_name :

    Mean Tb in cold core + cold anvil area

    units :

    K

    [400 values with dtype=float32]

  • core_meantb
    (times)
    float32
    ...

    long_name :

    Mean Tb in cold core area

    units :

    K

    [400 values with dtype=float32]

  • lat_mintb
    (times)
    float32
    ...

    long_name :

    Latitude with min Tb

    units :

    degree

    [400 values with dtype=float32]

  • lon_mintb
    (times)
    float32
    ...

    long_name :

    Longitude with min Tb

    units :

    degree

    [400 values with dtype=float32]

  • mcs_duration
    ()
    int64
    ...

    long_name :

    Duration of MCS stage

    units :

    unitless

    comments :

    Multiply by time_resolution_hour to convert to physical units

    [1 values with dtype=int64]

  • mcs_status
    (times)
    float32
    ...

    long_name :

    Flag indicating the status of MCS based on Tb. 1 = Yes, 0 = No

    units :

    unitless

    [400 values with dtype=float32]

  • ccs_area
    (times)
    float32
    6.2e+03 1.39e+04 ... nan nan

    long_name :

    Area of cold cloud shield

    units :

    km^2

    array([ 6200., 13900., 17700., ...,    nan,    nan,    nan], dtype=float32)

  • merge_cloudnumber
    (times, mergers)
    float64
    ...

    long_name :

    Cloud numbers that merge into MCS

    units :

    unitless

    [20000 values with dtype=float64]

  • split_cloudnumber
    (times, mergers)
    float64
    ...

    long_name :

    Cloud numbers that split from MCS

    units :

    unitless

    [20000 values with dtype=float64]

  • merge_ccs_area
    (times, mergers)
    float32
    ...

    long_name :

    Cold cloud shield area that merge into MCS

    units :

    km^2

    [20000 values with dtype=float32]

  • split_ccs_area
    (times, mergers)
    float32
    ...

    long_name :

    Cold cloud shield area that split from MCS

    units :

    km^2

    [20000 values with dtype=float32]

  • pf_npf
    (times)
    float32
    ...

    long_name :

    Number of PF in the cloud

    units :

    unitless

    [400 values with dtype=float32]

  • pf_lon
    (times, nmaxpf)
    float32
    ...

    long_name :

    Mean longitude of PF

    units :

    degrees

    [1200 values with dtype=float32]

  • pf_lat
    (times, nmaxpf)
    float32
    ...

    long_name :

    Mean latitude of PF

    units :

    degrees

    [1200 values with dtype=float32]

  • pf_area
    (times, nmaxpf)
    float32
    ...

    long_name :

    Area of PF

    units :

    km^2

    [1200 values with dtype=float32]

  • pf_rainrate
    (times, nmaxpf)
    float32
    ...

    long_name :

    Mean rain rate of PF

    units :

    mm/h

    [1200 values with dtype=float32]

  • pf_skewness
    (times, nmaxpf)
    float32
    ...

    long_name :

    Rain rate skewness of PF

    units :

    unitless

    [1200 values with dtype=float32]

  • pf_majoraxis
    (times, nmaxpf)
    float32
    ...

    long_name :

    Major axis length of PF

    units :

    km

    [1200 values with dtype=float32]

  • pf_minoraxis
    (times, nmaxpf)
    float32
    ...

    long_name :

    Minor axis length of PF

    units :

    km

    [1200 values with dtype=float32]

  • pf_aspectratio
    (times, nmaxpf)
    float32
    ...

    long_name :

    Aspect ratio (major axis / minor axis) of PF

    units :

    unitless

    [1200 values with dtype=float32]

  • pf_orientation
    (times, nmaxpf)
    float32
    ...

    long_name :

    Orientation of major axis of PF

    units :

    degrees

    [1200 values with dtype=float32]

  • pf_perimeter
    (times, nmaxpf)
    float32
    ...

    long_name :

    Perimeter of PF

    units :

    km

    [1200 values with dtype=float32]

  • pf_eccentricity
    (times, nmaxpf)
    float32
    ...

    long_name :

    Eccentricity of PF

    units :

    unitless

    [1200 values with dtype=float32]

  • pf_lon_centroid
    (times, nmaxpf)
    float32
    ...

    long_name :

    Centroid longitude of PF

    units :

    degrees

    [1200 values with dtype=float32]

  • pf_lat_centroid
    (times, nmaxpf)
    float32
    ...

    long_name :

    Centroid latitude of PF

    units :

    degrees

    [1200 values with dtype=float32]

  • pf_lon_weightedcentroid
    (times, nmaxpf)
    float32
    ...

    long_name :

    Weighted centroid longitude of PF

    units :

    degrees

    [1200 values with dtype=float32]

  • pf_lat_weightedcentroid
    (times, nmaxpf)
    float32
    ...

    long_name :

    Weighted centroid latitude of PF

    units :

    degrees

    [1200 values with dtype=float32]

  • pf_lon_maxrainrate
    (times, nmaxpf)
    float32
    ...

    long_name :

    Longitude with max rain rate

    units :

    degree

    [1200 values with dtype=float32]

  • pf_lat_maxrainrate
    (times, nmaxpf)
    float32
    ...

    long_name :

    Latitude with max rain rate

    units :

    degree

    [1200 values with dtype=float32]

  • pf_maxrainrate
    (times, nmaxpf)
    float32
    ...

    long_name :

    Max rain rate of PF

    units :

    mm/h

    [1200 values with dtype=float32]

  • pf_accumrain
    (times, nmaxpf)
    float32
    ...

    long_name :

    Accumulate precipitation of PF

    units :

    mm/h

    [1200 values with dtype=float32]

  • pf_accumrainheavy
    (times, nmaxpf)
    float32
    ...

    long_name :

    Accumulated heavy precipitation of PF

    units :

    mm/h

    heavy_rainrate_threshold :

    10.0

    [1200 values with dtype=float32]

  • pf_landfrac
    (times)
    float32
    ...

    long_name :

    Fraction of PF over land

    units :

    fraction

    [400 values with dtype=float32]

  • total_rain
    (times)
    float32
    ...

    long_name :

    Total precipitation under cold cloud shield (all rainfall included)

    units :

    mm/h

    [400 values with dtype=float32]

  • total_heavyrain
    (times)
    float32
    ...

    long_name :

    Total heavy precipitation under cold cloud shield

    units :

    mm/h

    heavy_rainrate_threshold :

    10.0

    [400 values with dtype=float32]

  • rainrate_heavyrain
    (times)
    float32
    ...

    long_name :

    Mean heavy rain rate under cold cloud shield

    units :

    mm/h

    heavy_rainrate_threshold :

    10.0

    [400 values with dtype=float32]

  • pf_lifetime
    ()
    float32
    ...

    long_name :

    MCS lifetime when a significant PF is present

    units :

    hour

    [1 values with dtype=float32]

  • pf_mcsstatus
    (times)
    float64
    ...

    long_name :

    Flag indicating the status of MCS based on PF. 1 = Yes, 0 = No

    units :

    unitless

    [400 values with dtype=float64]

  • movement_distance
    (times)
    float32
    ...

    long_name :

    Movement distance along angle movement_theta

    units :

    km

    lag :

    1

    comments :

    This is the total movement along the angle theta between lag estimates.

    [400 values with dtype=float32]

  • movement_speed
    (times)
    float32
    ...

    long_name :

    Movement speed along angle movement_theta

    units :

    m/s

    [400 values with dtype=float32]

  • movement_theta
    (times)
    float32
    ...

    long_name :

    Movement direction

    units :

    degrees

    comments :

    0:northward, 90:eastward, 180:southward, 270:westward.

    [400 values with dtype=float32]

  • movement_distance_x
    (times)
    float32
    ...

    long_name :

    East-West component of movement distance

    units :

    km

    [400 values with dtype=float32]

  • movement_distance_y
    (times)
    float32
    ...

    long_name :

    North-South component of movement distance

    units :

    km

    [400 values with dtype=float32]

• Indexes: (3)
  • times
    PandasIndex

    PandasIndex(Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
           ...
           390, 391, 392, 393, 394, 395, 396, 397, 398, 399],
          dtype='int64', name='times', length=400))

  • mergers
    PandasIndex

    PandasIndex(Index([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16, 17,
           18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35,
           36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49],
          dtype='int64', name='mergers'))

  • nmaxpf
    PandasIndex

    PandasIndex(Index([0, 1, 2], dtype='int64', name='nmaxpf'))

• Attributes: (21)

  Title :

  Statistics of each MCS track

  Institution :

  Pacific Northwest National Laboratory

  Contact :

  Zhe Feng, zhe.feng@pnnl.gov

  Created_on :

  Sun Apr 13 21:03:31 2025

  startdate :

  20190901.0000

  enddate :

  20200901.0000

  timegap_hour :

  3.1

  time_resolution_hour :

  1.0

  pixel_radius_km :

  10.0

  MCS_duration_hr :

  4

  MCS_area_km^2 :

  40000

  nmaxpf :

  3

  PF_rainrate_thresh :

  2.0

  heavy_rainrate_thresh :

  10.0

  landfrac_thresh :

  [ 0 90]

  MCS_PF_majoraxis_thresh :

  0

  MCS_PF_duration_thresh :

  4

  PF_PF_min_majoraxis_thresh :

  20

  mcs_min_rainvol_thresh :

  20000

  mcs_volrain_duration_thresh :

  1.0

  max_speed_thresh :

  50

# Access some of the track's properties:
# Scalars can be accessed using:
duration = track.track_duration.values.item()  # `.values.item()` unpacks the value into an int (in this case).
print(f'track_duration: {duration}')
# For times, it is useful to convert from a np.datetime64[ns] to a pandas Timestamp object or native datetime object:
start_basetime = pd.Timestamp(track.start_basetime.values.item()).to_pydatetime()
print(f'start_basetime: {start_basetime}')

# Lifetime values can be accessed:
# Note, e.g. area values are nan after the duration of the track:
print(f'area (full): {track.area.values}')
# Slice based on duration:
print(f'area (sliced): {track.area.values[:duration]}')

track_duration: 27
start_basetime: 2019-09-17 01:00:00
area (full): [7.4000e+03 1.3900e+04 1.7700e+04 5.5400e+04 9.1900e+04 1.4080e+05
 1.7750e+05 2.1500e+05 2.8800e+05 3.5320e+05 4.0140e+05 4.5130e+05
 5.0420e+05 6.4920e+05 6.2250e+05 1.0178e+06 1.4145e+06 1.0212e+06
 1.4063e+06 9.5510e+05 7.8290e+05 9.6880e+05 5.4550e+05 5.0100e+05
 4.5750e+05 1.9970e+05 1.3000e+03        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan        nan        nan
        nan        nan        nan        nan]
area (sliced): [7.4000e+03 1.3900e+04 1.7700e+04 5.5400e+04 9.1900e+04 1.4080e+05
 1.7750e+05 2.1500e+05 2.8800e+05 3.5320e+05 4.0140e+05 4.5130e+05
 5.0420e+05 6.4920e+05 6.2250e+05 1.0178e+06 1.4145e+06 1.0212e+06
 1.4063e+06 9.5510e+05 7.8290e+05 9.6880e+05 5.4550e+05 5.0100e+05
 4.5750e+05 1.9970e+05 1.3000e+03]

# Plot the MCS cloud shield area time evolution:
track.ccs_area.plot()

[<matplotlib.lines.Line2D at 0x7f7d9defdf10>]

# Plot multiple variables for this track

fig, ax1 = plt.subplots(figsize=(10, 5))
track.ccs_area.plot.line(ax=ax1, label='CCS Area')

ax2 = ax1.twinx()
track.corecold_mintb.plot(ax=ax2, color='green', label='Min Tb')
ax2.spines['right'].set_position(('outward', 60))  # Move the third y-axis to the right
ax2.invert_yaxis()

ax3 = ax1.twinx()
track.pf_maxrainrate.isel(nmaxpf=0).plot(ax=ax3, color='orange', label='Max Rain Rate')
ax3.set_title('')

ax4 = ax1.twinx()
track.total_rain.plot(ax=ax4, color='red', label='Total Rain')
ax4.spines['right'].set_position(('outward', 120))  # Move the third y-axis to the right

fig.legend()

<matplotlib.legend.Legend at 0x7f7d9dd69820>

# mergers data are 2D fields (with -9999 indicating no values for merger number N):
track.merge_cloudnumber.values[:duration, :]

array([[nan, nan, nan, ..., nan, nan, nan],
       [nan, nan, nan, ..., nan, nan, nan],
       [nan, nan, nan, ..., nan, nan, nan],
       ...,
       [nan, nan, nan, ..., nan, nan, nan],
       [nan, nan, nan, ..., nan, nan, nan],
       [nan, nan, nan, ..., nan, nan, nan]])

# Similarly for PF data (with nan indicating no value):
track.pf_rainrate.values[:duration, :]

array([[17.444798 ,        nan,        nan],
       [19.396593 ,        nan,        nan],
       [ 3.642234 ,        nan,        nan],
       [14.076908 , 16.538572 , 12.292864 ],
       [13.964077 , 18.771038 ,  3.4211912],
       [13.651215 , 14.239631 ,  4.3471093],
       [13.093377 ,  9.806874 , 20.726156 ],
       [14.3507805, 17.64512  ,  2.9985306],
       [12.434007 , 19.540384 , 12.185244 ],
       [12.988246 , 12.031208 , 15.784163 ],
       [12.270396 , 11.649001 , 12.509272 ],
       [10.951215 ,  8.072648 ,  9.52699  ],
       [ 9.013996 ,  6.1185   , 11.534578 ],
       [11.8789425, 10.324547 ,  6.064827 ],
       [10.708405 ,  7.544305 , 13.022357 ],
       [ 5.462044 ,  7.9311924,  7.762195 ],
       [13.009703 ,  9.892664 , 11.157013 ],
       [ 9.685159 ,  5.8382173,  9.085187 ],
       [10.451695 , 12.007852 ,  6.4264364],
       [10.569543 ,  7.5295744,  8.378777 ],
       [ 7.369762 ,  9.8796625,  9.1030855],
       [ 7.2170215,  7.7586746,  7.728683 ],
       [ 5.010368 ,  5.5169315,  6.242524 ],
       [ 3.9187026,  4.519898 ,  5.0759807],
       [ 8.638662 ,  2.550455 ,  2.8447366],
       [       nan,        nan,        nan],
       [       nan,        nan,        nan]], dtype=float32)

# A simple plot of the track's position can be made using:
plt.scatter(track.meanlon.values[0], track.meanlat.values[0], marker='o', color='k')  # Start point.
plt.scatter(track.meanlon.values[duration - 1], track.meanlat.values[duration - 1], marker='x', color='darkorange')  # End point.
plt.plot(track.meanlon.values, track.meanlat.values)

[<matplotlib.lines.Line2D at 0x7f7d9ddee1e0>]

def add_ccs_area_swath(ax, track, n_points=20):
    """Adds an area swath of the cold cloud system area, treating each CCS as a circle."""
    try:
        # N.B. these are optional dependencies.
        import shapely.geometry
        import shapely.ops
    except ImportError:
        print('shapely not installed')
        return
    duration = track.track_duration.values.item()
    time_indices = range(duration)
    
    # geoms will contain all the circles.
    geoms = []
    for i in time_indices:
        lon = track.meanlon.values[i]
        lat = track.meanlat.values[i]
        radius = np.sqrt(track.ccs_area.values[i] / np.pi) * 1e3
        circle_points = cartopy.geodesic.Geodesic().circle(
            lon=lon, lat=lat, radius=radius, n_samples=n_points, endpoint=False
        )
        geom = shapely.geometry.Polygon(circle_points)
        geoms.append(geom)
    # Combine all the circles into a CCS swath.
    full_geom = shapely.ops.unary_union(geoms)
    ax.add_geometries(
        (full_geom,),
        crs=cartopy.crs.PlateCarree(),
        facecolor="none",
        edgecolor="royalblue",
        linewidth=2,
    )

extent = track.meanlon.min().item() - 10, track.meanlon.max().item() + 10, track.meanlat.min().item() - 10, track.meanlat.max().item() + 10

# Nicer figure using cartopy projections and showing a circle based on the CCS area.
fig, ax = plt.subplots(subplot_kw={'projection': ccrs.PlateCarree()})
fig.set_size_inches((10, 5))
ax.coastlines()
ax.add_feature(cf.BORDERS, linewidth=0.4)
ax.scatter(track.meanlon.values[0], track.meanlat.values[0], marker='o', color='k')
ax.scatter(track.meanlon.values[duration - 1], track.meanlat.values[duration - 1], marker='x', color='darkorange')
ax.plot(track.meanlon.values, track.meanlat.values)
add_ccs_area_swath(ax, track)
extent = track.meanlon.min().item() - 20, track.meanlon.max().item() + 20, track.meanlat.min().item() - 20, track.meanlat.max().item() + 20
ax.set_extent(extent)

Group properties of tracks

Group properties for lots of tracks can be easily calculated by accessing the fields on an xarray.Dataset that contains many tracks.

# mean track duration:
dstracks.track_duration.values.mean()

np.float64(18.04881699384031)

# Same thing using xarray:
dstracks.track_duration.mean().values.item()

18.04881699384031

# Tropical duration:
dstracks_tropical.track_duration.values.mean()

np.float64(18.115322698668226)

# Mean area for each track:
# `np.nanX` functions are useful for naturally dealing with the missing data.
mean_areas = np.nanmean(dstracks.ccs_area.values, axis=1)

plt.hist(mean_areas, bins=np.linspace(0, 1e6, 101))

(array([0.000e+00, 2.500e+01, 1.871e+03, 6.303e+03, 6.692e+03, 5.703e+03,
        4.438e+03, 3.576e+03, 2.807e+03, 2.204e+03, 1.860e+03, 1.530e+03,
        1.381e+03, 1.139e+03, 9.780e+02, 8.390e+02, 8.210e+02, 6.660e+02,
        6.260e+02, 5.390e+02, 4.520e+02, 4.470e+02, 4.320e+02, 3.830e+02,
        3.480e+02, 3.010e+02, 2.980e+02, 2.560e+02, 2.490e+02, 2.530e+02,
        2.150e+02, 1.950e+02, 1.890e+02, 1.810e+02, 1.600e+02, 1.550e+02,
        1.420e+02, 1.580e+02, 1.270e+02, 1.300e+02, 1.280e+02, 1.180e+02,
        9.800e+01, 9.200e+01, 9.100e+01, 1.040e+02, 1.050e+02, 8.000e+01,
        8.700e+01, 8.600e+01, 9.400e+01, 6.500e+01, 7.600e+01, 6.500e+01,
        4.800e+01, 6.100e+01, 6.400e+01, 5.100e+01, 5.500e+01, 6.300e+01,
        5.000e+01, 6.000e+01, 2.900e+01, 4.900e+01, 4.400e+01, 3.900e+01,
        4.900e+01, 4.100e+01, 4.200e+01, 3.300e+01, 3.500e+01, 2.700e+01,
        2.800e+01, 3.000e+01, 3.900e+01, 3.000e+01, 3.900e+01, 2.200e+01,
        3.000e+01, 2.300e+01, 3.300e+01, 1.500e+01, 2.700e+01, 2.700e+01,
        2.100e+01, 2.800e+01, 2.100e+01, 1.100e+01, 2.200e+01, 2.200e+01,
        2.100e+01, 1.300e+01, 1.200e+01, 5.000e+00, 2.500e+01, 1.700e+01,
        1.000e+01, 1.200e+01, 1.300e+01, 2.300e+01]),
 array([      0.,   10000.,   20000.,   30000.,   40000.,   50000.,
          60000.,   70000.,   80000.,   90000.,  100000.,  110000.,
         120000.,  130000.,  140000.,  150000.,  160000.,  170000.,
         180000.,  190000.,  200000.,  210000.,  220000.,  230000.,
         240000.,  250000.,  260000.,  270000.,  280000.,  290000.,
         300000.,  310000.,  320000.,  330000.,  340000.,  350000.,
         360000.,  370000.,  380000.,  390000.,  400000.,  410000.,
         420000.,  430000.,  440000.,  450000.,  460000.,  470000.,
         480000.,  490000.,  500000.,  510000.,  520000.,  530000.,
         540000.,  550000.,  560000.,  570000.,  580000.,  590000.,
         600000.,  610000.,  620000.,  630000.,  640000.,  650000.,
         660000.,  670000.,  680000.,  690000.,  700000.,  710000.,
         720000.,  730000.,  740000.,  750000.,  760000.,  770000.,
         780000.,  790000.,  800000.,  810000.,  820000.,  830000.,
         840000.,  850000.,  860000.,  870000.,  880000.,  890000.,
         900000.,  910000.,  920000.,  930000.,  940000.,  950000.,
         960000.,  970000.,  980000.,  990000., 1000000.]),
 <BarContainer object of 100 artists>)

# Same thing using xarray:
xr_mean_areas = dstracks.ccs_area.mean(dim='times', skipna=True).values

(mean_areas == xr_mean_areas).all()

np.True_

Accessing pixel-level data

Pixel-level data is stored separately from the MCS tracks data. These files contain the raw brightness temperature (tb) and precipitation fields, as well as a derived cloudnumber field, which is the area covered by the CCS and can be linked to a given track (see below).

# Read MCS mask data:
dsmask = xr.open_dataset(
    maskdir,
    chunks={},
    # mask_and_scale=False,
)
dsmask

<xarray.Dataset> Size: 219GB
Dimensions:   (cell: 3145728, time: 8713)
Coordinates:
  * cell      (cell) int64 25MB 0 1 2 3 4 ... 3145724 3145725 3145726 3145727
    lat       (cell) float64 25MB dask.array<chunksize=(262144,), meta=np.ndarray>
    lon       (cell) float64 25MB dask.array<chunksize=(262144,), meta=np.ndarray>
  * time      (time) datetime64[ns] 70kB 2019-09-01 ... 2020-09-01
Data variables:
    mcs_mask  (time, cell) float64 219GB dask.array<chunksize=(24, 262144), meta=np.ndarray>

xarray.Dataset

• Dimensions:
  • cell: 3145728
  • time: 8713
• Coordinates: (4)
  • cell
    (cell)
    int64
    0 1 2 3 ... 3145725 3145726 3145727

    array([      0,       1,       2, ..., 3145725, 3145726, 3145727])

  • lat
    (cell)
    float64
    dask.array<chunksize=(262144,), meta=np.ndarray>

    Array Chunk Bytes 24.00 MiB 2.00 MiB Shape (3145728,) (262144,) Dask graph 12 chunks in 2 graph layers Data type float64 numpy.ndarray

  • lon
    (cell)
    float64
    dask.array<chunksize=(262144,), meta=np.ndarray>

    Array Chunk Bytes 24.00 MiB 2.00 MiB Shape (3145728,) (262144,) Dask graph 12 chunks in 2 graph layers Data type float64 numpy.ndarray

  • time
    (time)
    datetime64[ns]
    2019-09-01 ... 2020-09-01

    long_name :

    Base time in Epoch

    array(['2019-09-01T00:00:00.000000000', '2019-09-01T01:00:00.000000000',
           '2019-09-01T02:00:00.000000000', ..., '2020-08-31T22:00:00.000000000',
           '2020-08-31T23:00:00.000000000', '2020-09-01T00:00:00.000000000'],
          dtype='datetime64[ns]')

• Data variables: (1)
  • mcs_mask
    (time, cell)
    float64
    dask.array<chunksize=(24, 262144), meta=np.ndarray>

    Array Chunk Bytes 204.21 GiB 48.00 MiB Shape (8713, 3145728) (24, 262144) Dask graph 4368 chunks in 2 graph layers Data type float64 numpy.ndarray

• Indexes: (2)
  • cell
    PandasIndex

    PandasIndex(Index([      0,       1,       2,       3,       4,       5,       6,       7,
                 8,       9,
           ...
           3145718, 3145719, 3145720, 3145721, 3145722, 3145723, 3145724, 3145725,
           3145726, 3145727],
          dtype='int64', name='cell', length=3145728))

  • time
    PandasIndex

    PandasIndex(DatetimeIndex(['2019-09-01 00:00:00', '2019-09-01 01:00:00',
                   '2019-09-01 02:00:00', '2019-09-01 03:00:00',
                   '2019-09-01 04:00:00', '2019-09-01 05:00:00',
                   '2019-09-01 06:00:00', '2019-09-01 07:00:00',
                   '2019-09-01 08:00:00', '2019-09-01 09:00:00',
                   ...
                   '2020-08-31 15:00:00', '2020-08-31 16:00:00',
                   '2020-08-31 17:00:00', '2020-08-31 18:00:00',
                   '2020-08-31 19:00:00', '2020-08-31 20:00:00',
                   '2020-08-31 21:00:00', '2020-08-31 22:00:00',
                   '2020-08-31 23:00:00', '2020-09-01 00:00:00'],
                  dtype='datetime64[ns]', name='time', length=8713, freq=None))

• Attributes: (0)

# Read OLR data
dsolr = xr.open_dataset(
    olrdir,
    chunks={},
    # mask_and_scale=False,
)
dsolr

<xarray.Dataset> Size: 119GB
Dimensions:  (crs: 0, time: 9457, cell: 3145728)
Coordinates:
  * crs      (crs) float64 0B 
  * time     (time) object 76kB 2019-08-01 00:00:00 ... 2020-09-01 00:00:00
Dimensions without coordinates: cell
Data variables:
    rlut     (time, cell) float32 119GB dask.array<chunksize=(24, 4096), meta=np.ndarray>
Attributes: (12/22)
    Conventions:                  CF-1.8
    atm_initial_conditions_file:  /lustre/orion/cli115/world-shared/e3sm/inpu...
    averaging_frequency:          1
    averaging_frequency_units:    nhours
    averaging_type:               INSTANT
    case:                         cess-control.ne1024pg2_ne1024pg2.F2010-SCRE...
    ...                           ...
    product:                      model-output
    realm:                        atmos
    run_t0:                       2019-08-01-00000
    source:                       E3SM Atmosphere Model (EAMxx)
    topography_file:              /lustre/orion/cli115/world-shared/e3sm/inpu...
    username:                     noel

xarray.Dataset

• Dimensions:
  • crs: 0
  • time: 9457
  • cell: 3145728
• Coordinates: (2)
  • crs
    (crs)
    float64

    grid_mapping_name :

    healpix

    healpix_nside :

    512

    healpix_order :

    nest

    array([], dtype=float64)

  • time
    (time)
    object
    2019-08-01 00:00:00 ... 2020-09-...

    long_name :

    time

    array([cftime.DatetimeNoLeap(2019, 8, 1, 0, 0, 0, 0, has_year_zero=True),
           cftime.DatetimeNoLeap(2019, 8, 1, 1, 0, 0, 0, has_year_zero=True),
           cftime.DatetimeNoLeap(2019, 8, 1, 2, 0, 0, 0, has_year_zero=True), ...,
           cftime.DatetimeNoLeap(2020, 8, 31, 22, 0, 0, 0, has_year_zero=True),
           cftime.DatetimeNoLeap(2020, 8, 31, 23, 0, 0, 0, has_year_zero=True),
           cftime.DatetimeNoLeap(2020, 9, 1, 0, 0, 0, 0, has_year_zero=True)],
          dtype=object)

• Data variables: (1)
  • rlut
    (time, cell)
    float32
    dask.array<chunksize=(24, 4096), meta=np.ndarray>

    long_name :

    LW_flux_up_at_model_top

    units :

    W/m2

    Array Chunk Bytes 110.82 GiB 384.00 kiB Shape (9457, 3145728) (24, 4096) Dask graph 303360 chunks in 2 graph layers Data type float32 numpy.ndarray

• Indexes: (2)
  • crs
    PandasIndex

    PandasIndex(Index([], dtype='float64', name='crs'))

  • time
    PandasIndex

    PandasIndex(CFTimeIndex([2019-08-01 00:00:00, 2019-08-01 01:00:00, 2019-08-01 02:00:00,
                 2019-08-01 03:00:00, 2019-08-01 04:00:00, 2019-08-01 05:00:00,
                 2019-08-01 06:00:00, 2019-08-01 07:00:00, 2019-08-01 08:00:00,
                 2019-08-01 09:00:00,
                 ...
                 2020-08-31 15:00:00, 2020-08-31 16:00:00, 2020-08-31 17:00:00,
                 2020-08-31 18:00:00, 2020-08-31 19:00:00, 2020-08-31 20:00:00,
                 2020-08-31 21:00:00, 2020-08-31 22:00:00, 2020-08-31 23:00:00,
                 2020-09-01 00:00:00],
                dtype='object', length=9457, calendar='noleap', freq=None))

• Attributes: (22)

  Conventions :

  CF-1.8

  atm_initial_conditions_file :

  /lustre/orion/cli115/world-shared/e3sm/inputdata/atm/scream/init/screami_ne1024np4L128_era5-20190801-topoadjx6t_20230620.nc

  averaging_frequency :

  1

  averaging_frequency_units :

  nhours

  averaging_type :

  INSTANT

  case :

  cess-control.ne1024pg2_ne1024pg2.F2010-SCREAMv1.cess-oct2

  case_t0 :

  2019-08-01-00000

  contact :

  e3sm-data-support@llnl.gov

  eamxx_version :

  1.0.0

  fp_precision :

  single

  git_version :

  f4f0672bb4283cad9347808e9576859974fffb12

  history :

  created on Sun Oct 8 17:20:09 2023

  hostname :

  frontier-scream-gpu

  institution_id :

  E3SM-Projet

  institutions :

  LLNL (Lawrence Livermore National Laboratory, Livermore, CA 94550, USA); ANL (Argonne National Laboratory, Argonne, IL 60439, USA); BNL (Brookhaven National Laboratory, Upton, NY 11973, USA); LANL (Los Alamos National Laboratory, Los Alamos, NM 87545, USA); LBNL (Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA); ORNL (Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA); PNNL (Pacific Northwest National Laboratory, Richland, WA 99352, USA); SNL (Sandia National Laboratories, Albuquerque, NM 87185, USA). Mailing address: LLNL Climate Program, c/o David C. Bader, Principal Investigator, L-103, 7000 East Avenue, Livermore, CA 94550, USA

  max_snapshots_per_file :

  96

  product :

  model-output

  realm :

  atmos

  run_t0 :

  2019-08-01-00000

  source :

  E3SM Atmosphere Model (EAMxx)

  topography_file :

  /lustre/orion/cli115/world-shared/e3sm/inputdata/atm/cam/topo/USGS-gtopo30_ne1024np4pg2_x6t-SGH.c20210614.nc

  username :

  noel

# Read precipitation data
dspr = xr.open_dataset(
    prdir,
    chunks={},
    # mask_and_scale=False,
)
dspr

<xarray.Dataset> Size: 119GB
Dimensions:  (crs: 0, time: 9457, cell: 3145728)
Coordinates:
  * crs      (crs) float64 0B 
  * time     (time) object 76kB 2019-08-01 00:00:00 ... 2020-09-01 00:00:00
Dimensions without coordinates: cell
Data variables:
    pr       (time, cell) float32 119GB dask.array<chunksize=(48, 4096), meta=np.ndarray>
Attributes: (12/22)
    Conventions:                  CF-1.8
    atm_initial_conditions_file:  /lustre/orion/cli115/world-shared/e3sm/inpu...
    averaging_frequency:          1
    averaging_frequency_units:    nhours
    averaging_type:               INSTANT
    case:                         cess-control.ne1024pg2_ne1024pg2.F2010-SCRE...
    ...                           ...
    product:                      model-output
    realm:                        atmos
    run_t0:                       2019-08-01-00000
    source:                       E3SM Atmosphere Model (EAMxx)
    topography_file:              /lustre/orion/cli115/world-shared/e3sm/inpu...
    username:                     noel

xarray.Dataset

• Dimensions:
  • crs: 0
  • time: 9457
  • cell: 3145728
• Coordinates: (2)
  • crs
    (crs)
    float64

    grid_mapping_name :

    healpix

    healpix_nside :

    512

    healpix_order :

    nest

    array([], dtype=float64)

  • time
    (time)
    object
    2019-08-01 00:00:00 ... 2020-09-...

    long_name :

    time

    array([cftime.DatetimeNoLeap(2019, 8, 1, 0, 0, 0, 0, has_year_zero=True),
           cftime.DatetimeNoLeap(2019, 8, 1, 1, 0, 0, 0, has_year_zero=True),
           cftime.DatetimeNoLeap(2019, 8, 1, 2, 0, 0, 0, has_year_zero=True), ...,
           cftime.DatetimeNoLeap(2020, 8, 31, 22, 0, 0, 0, has_year_zero=True),
           cftime.DatetimeNoLeap(2020, 8, 31, 23, 0, 0, 0, has_year_zero=True),
           cftime.DatetimeNoLeap(2020, 9, 1, 0, 0, 0, 0, has_year_zero=True)],
          dtype=object)

• Data variables: (1)
  • pr
    (time, cell)
    float32
    dask.array<chunksize=(48, 4096), meta=np.ndarray>

    long_name :

    precip_total_surf_mass_flux

    units :

    m s^-1

    Array Chunk Bytes 110.82 GiB 768.00 kiB Shape (9457, 3145728) (48, 4096) Dask graph 152064 chunks in 2 graph layers Data type float32 numpy.ndarray

• Indexes: (2)
  • crs
    PandasIndex

    PandasIndex(Index([], dtype='float64', name='crs'))

  • time
    PandasIndex

    PandasIndex(CFTimeIndex([2019-08-01 00:00:00, 2019-08-01 01:00:00, 2019-08-01 02:00:00,
                 2019-08-01 03:00:00, 2019-08-01 04:00:00, 2019-08-01 05:00:00,
                 2019-08-01 06:00:00, 2019-08-01 07:00:00, 2019-08-01 08:00:00,
                 2019-08-01 09:00:00,
                 ...
                 2020-08-31 15:00:00, 2020-08-31 16:00:00, 2020-08-31 17:00:00,
                 2020-08-31 18:00:00, 2020-08-31 19:00:00, 2020-08-31 20:00:00,
                 2020-08-31 21:00:00, 2020-08-31 22:00:00, 2020-08-31 23:00:00,
                 2020-09-01 00:00:00],
                dtype='object', length=9457, calendar='noleap', freq=None))

• Attributes: (22)

  Conventions :

  CF-1.8

  atm_initial_conditions_file :

  /lustre/orion/cli115/world-shared/e3sm/inputdata/atm/scream/init/screami_ne1024np4L128_era5-20190801-topoadjx6t_20230620.nc

  averaging_frequency :

  1

  averaging_frequency_units :

  nhours

  averaging_type :

  INSTANT

  case :

  cess-control.ne1024pg2_ne1024pg2.F2010-SCREAMv1.cess-oct2

  case_t0 :

  2019-08-01-00000

  contact :

  e3sm-data-support@llnl.gov

  eamxx_version :

  1.0.0

  fp_precision :

  single

  git_version :

  f4f0672bb4283cad9347808e9576859974fffb12

  history :

  created on Sun Oct 8 17:20:09 2023

  hostname :

  frontier-scream-gpu

  institution_id :

  E3SM-Projet

  institutions :

  LLNL (Lawrence Livermore National Laboratory, Livermore, CA 94550, USA); ANL (Argonne National Laboratory, Argonne, IL 60439, USA); BNL (Brookhaven National Laboratory, Upton, NY 11973, USA); LANL (Los Alamos National Laboratory, Los Alamos, NM 87545, USA); LBNL (Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA); ORNL (Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA); PNNL (Pacific Northwest National Laboratory, Richland, WA 99352, USA); SNL (Sandia National Laboratories, Albuquerque, NM 87185, USA). Mailing address: LLNL Climate Program, c/o David C. Bader, Principal Investigator, L-103, 7000 East Avenue, Livermore, CA 94550, USA

  max_snapshots_per_file :

  96

  product :

  model-output

  realm :

  atmos

  run_t0 :

  2019-08-01-00000

  source :

  E3SM Atmosphere Model (EAMxx)

  topography_file :

  /lustre/orion/cli115/world-shared/e3sm/inputdata/atm/cam/topo/USGS-gtopo30_ne1024np4pg2_x6t-SGH.c20210614.nc

  username :

  noel

# Convert time components from datasets to deal with different calendars
olr_times = [(t.year, t.month, t.day, t.hour) for t in dsolr.time.values]
mask_times = [(pd.Timestamp(t).year, pd.Timestamp(t).month, 
               pd.Timestamp(t).day, pd.Timestamp(t).hour) 
              for t in dsmask.time.values]

# Find matching indices
matching_indices = [i for i, t in enumerate(olr_times) if t in set(mask_times)]

# Select using indices
dsolr_subset = dsolr.isel(time=matching_indices)
# Precipitation data should have identical times with OLR data
dspr_subset = dspr.isel(time=matching_indices)

# Verify number of timestamps
print(f"Original dsolr time count: {dsolr.time.size}")
print(f"Matching times found: {len(matching_indices)}")
print(f"dsmask time count: {dsmask.time.size}")

Original dsolr time count: 9457
Matching times found: 8713
dsmask time count: 8713

# Create a new dsmask with the time coordinate from dsolr_subset
dsmask_ = dsmask.copy()

# Replace the time coordinate with the matching times from dsolr_subset
# This must maintain the same order as in the original dsmask
dsmask_ = dsmask_.assign_coords(
    time=dsolr_subset.time.values
)

# Combine the datasets
ds = xr.merge([dsolr_subset, dspr_subset, dsmask_], combine_attrs='drop_conflicts')

# Rechunk the dataset to match the mask
ds = ds.chunk(dict(dsmask_.chunks))
ds

<xarray.Dataset> Size: 439GB
Dimensions:   (crs: 0, time: 8713, cell: 3145728)
Coordinates:
  * crs       (crs) float64 0B 
  * time      (time) object 70kB 2019-09-01 00:00:00 ... 2020-09-01 00:00:00
  * cell      (cell) int64 25MB 0 1 2 3 4 ... 3145724 3145725 3145726 3145727
    lat       (cell) float64 25MB dask.array<chunksize=(262144,), meta=np.ndarray>
    lon       (cell) float64 25MB dask.array<chunksize=(262144,), meta=np.ndarray>
Data variables:
    rlut      (time, cell) float32 110GB dask.array<chunksize=(24, 262144), meta=np.ndarray>
    pr        (time, cell) float32 110GB dask.array<chunksize=(24, 262144), meta=np.ndarray>
    mcs_mask  (time, cell) float64 219GB dask.array<chunksize=(24, 262144), meta=np.ndarray>
Attributes: (12/22)
    Conventions:                  CF-1.8
    atm_initial_conditions_file:  /lustre/orion/cli115/world-shared/e3sm/inpu...
    averaging_frequency:          1
    averaging_frequency_units:    nhours
    averaging_type:               INSTANT
    case:                         cess-control.ne1024pg2_ne1024pg2.F2010-SCRE...
    ...                           ...
    product:                      model-output
    realm:                        atmos
    run_t0:                       2019-08-01-00000
    source:                       E3SM Atmosphere Model (EAMxx)
    topography_file:              /lustre/orion/cli115/world-shared/e3sm/inpu...
    username:                     noel

xarray.Dataset

• Dimensions:
  • crs: 0
  • time: 8713
  • cell: 3145728
• Coordinates: (5)
  • crs
    (crs)
    float64

    grid_mapping_name :

    healpix

    healpix_nside :

    512

    healpix_order :

    nest

    array([], dtype=float64)

  • time
    (time)
    object
    2019-09-01 00:00:00 ... 2020-09-...

    long_name :

    time

    array([cftime.DatetimeNoLeap(2019, 9, 1, 0, 0, 0, 0, has_year_zero=True),
           cftime.DatetimeNoLeap(2019, 9, 1, 1, 0, 0, 0, has_year_zero=True),
           cftime.DatetimeNoLeap(2019, 9, 1, 2, 0, 0, 0, has_year_zero=True), ...,
           cftime.DatetimeNoLeap(2020, 8, 31, 22, 0, 0, 0, has_year_zero=True),
           cftime.DatetimeNoLeap(2020, 8, 31, 23, 0, 0, 0, has_year_zero=True),
           cftime.DatetimeNoLeap(2020, 9, 1, 0, 0, 0, 0, has_year_zero=True)],
          dtype=object)

  • cell
    (cell)
    int64
    0 1 2 3 ... 3145725 3145726 3145727

    array([      0,       1,       2, ..., 3145725, 3145726, 3145727])

  • lat
    (cell)
    float64
    dask.array<chunksize=(262144,), meta=np.ndarray>

    Array Chunk Bytes 24.00 MiB 2.00 MiB Shape (3145728,) (262144,) Dask graph 12 chunks in 2 graph layers Data type float64 numpy.ndarray

  • lon
    (cell)
    float64
    dask.array<chunksize=(262144,), meta=np.ndarray>

    Array Chunk Bytes 24.00 MiB 2.00 MiB Shape (3145728,) (262144,) Dask graph 12 chunks in 2 graph layers Data type float64 numpy.ndarray

• Data variables: (3)
  • rlut
    (time, cell)
    float32
    dask.array<chunksize=(24, 262144), meta=np.ndarray>

    long_name :

    LW_flux_up_at_model_top

    units :

    W/m2

    Array Chunk Bytes 102.11 GiB 24.00 MiB Shape (8713, 3145728) (24, 262144) Dask graph 4368 chunks in 4 graph layers Data type float32 numpy.ndarray

  • pr
    (time, cell)
    float32
    dask.array<chunksize=(24, 262144), meta=np.ndarray>

    long_name :

    precip_total_surf_mass_flux

    units :

    m s^-1

    Array Chunk Bytes 102.11 GiB 24.00 MiB Shape (8713, 3145728) (24, 262144) Dask graph 4368 chunks in 4 graph layers Data type float32 numpy.ndarray

  • mcs_mask
    (time, cell)
    float64
    dask.array<chunksize=(24, 262144), meta=np.ndarray>

    Array Chunk Bytes 204.21 GiB 48.00 MiB Shape (8713, 3145728) (24, 262144) Dask graph 4368 chunks in 2 graph layers Data type float64 numpy.ndarray

• Indexes: (3)
  • crs
    PandasIndex

    PandasIndex(Index([], dtype='float64', name='crs'))

  • time
    PandasIndex

    PandasIndex(CFTimeIndex([2019-09-01 00:00:00, 2019-09-01 01:00:00, 2019-09-01 02:00:00,
                 2019-09-01 03:00:00, 2019-09-01 04:00:00, 2019-09-01 05:00:00,
                 2019-09-01 06:00:00, 2019-09-01 07:00:00, 2019-09-01 08:00:00,
                 2019-09-01 09:00:00,
                 ...
                 2020-08-31 15:00:00, 2020-08-31 16:00:00, 2020-08-31 17:00:00,
                 2020-08-31 18:00:00, 2020-08-31 19:00:00, 2020-08-31 20:00:00,
                 2020-08-31 21:00:00, 2020-08-31 22:00:00, 2020-08-31 23:00:00,
                 2020-09-01 00:00:00],
                dtype='object', length=8713, calendar='noleap', freq=None))

  • cell
    PandasIndex

    PandasIndex(Index([      0,       1,       2,       3,       4,       5,       6,       7,
                 8,       9,
           ...
           3145718, 3145719, 3145720, 3145721, 3145722, 3145723, 3145724, 3145725,
           3145726, 3145727],
          dtype='int64', name='cell', length=3145728))

• Attributes: (22)

  Conventions :

  CF-1.8

  atm_initial_conditions_file :

  /lustre/orion/cli115/world-shared/e3sm/inputdata/atm/scream/init/screami_ne1024np4L128_era5-20190801-topoadjx6t_20230620.nc

  averaging_frequency :

  1

  averaging_frequency_units :

  nhours

  averaging_type :

  INSTANT

  case :

  cess-control.ne1024pg2_ne1024pg2.F2010-SCREAMv1.cess-oct2

  case_t0 :

  2019-08-01-00000

  contact :

  e3sm-data-support@llnl.gov

  eamxx_version :

  1.0.0

  fp_precision :

  single

  git_version :

  f4f0672bb4283cad9347808e9576859974fffb12

  history :

  created on Sun Oct 8 17:20:09 2023

  hostname :

  frontier-scream-gpu

  institution_id :

  E3SM-Projet

  institutions :

  LLNL (Lawrence Livermore National Laboratory, Livermore, CA 94550, USA); ANL (Argonne National Laboratory, Argonne, IL 60439, USA); BNL (Brookhaven National Laboratory, Upton, NY 11973, USA); LANL (Los Alamos National Laboratory, Los Alamos, NM 87545, USA); LBNL (Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA); ORNL (Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA); PNNL (Pacific Northwest National Laboratory, Richland, WA 99352, USA); SNL (Sandia National Laboratories, Albuquerque, NM 87185, USA). Mailing address: LLNL Climate Program, c/o David C. Bader, Principal Investigator, L-103, 7000 East Avenue, Livermore, CA 94550, USA

  max_snapshots_per_file :

  96

  product :

  model-output

  realm :

  atmos

  run_t0 :

  2019-08-01-00000

  source :

  E3SM Atmosphere Model (EAMxx)

  topography_file :

  /lustre/orion/cli115/world-shared/e3sm/inputdata/atm/cam/topo/USGS-gtopo30_ne1024np4pg2_x6t-SGH.c20210614.nc

  username :

  noel

ds.rlut.attrs

{'long_name': 'LW_flux_up_at_model_top', 'units': 'W/m2'}

# Quick plot of global OLR
# im = egh.healpix_show(ds.isel(time=0).rlut, cmap='jet', dpi=150)
im = egh.healpix_show(ds.isel(time=0).rlut, vmin=80, vmax=300, cmap='Greys')
plt.colorbar(im, label=f"{ds.rlut.attrs['long_name']} ({ds.rlut.attrs['units']})", orientation='horizontal')
time_str = ds.time.isel(time=0).dt.strftime("%Y-%m-%d %H:%M").values.item()
plt.title(time_str)

Text(0.5, 1.0, '2019-09-01 00:00')

<Figure size 640x480 with 0 Axes>

ds.time.isel(time=0).dt.strftime("%Y-%m-%d %H:%M").values.item()

'2019-09-01 00:00'

# Quick plot of global precipitation
levels_pr = [0.5,1,2,4,8,16,32,64]
cmap = mpl.colormaps.get_cmap('jet')
norm = mpl.colors.BoundaryNorm(boundaries=levels_pr, ncolors=cmap.N)
# Precipitation
_pr = ds.isel(time=0).pr * 3600000  # Convert unit to mm/hr
_pr = _pr.where(_pr >= np.min(levels_pr), np.nan)

im = egh.healpix_show(_pr, norm=norm, cmap=cmap)
plt.colorbar(im, label=f"{ds.pr.attrs['long_name']} (mm/h)", orientation='horizontal')
time_str = ds.time.isel(time=0).dt.strftime("%Y-%m-%d %H:%M").values.item()
plt.title(time_str)

Text(0.5, 1.0, '2019-09-01 00:00')

<Figure size 640x480 with 0 Axes>

Linking tracks to pixel-level data

Every track has a corresponding track number in the pixel-level mask data. This can be used to link each track to its pixel-level data, using the timestamp and the track number. The timestamp can be used to determine which time step data to load, and the track number references the equivalent field in the HEALPix dataset (as shown in the figure above).

itrack = 2 # (Borneo)
# itrack = 17
# itrack = 40 # (Brazil)

# Select a tropical track from July 2019, that lasted more than buffer hours and less than 40.
dates = pd.DatetimeIndex(dstracks_tropical.start_basetime.values)  # These containers make it easy to select on year, month...
track = dstracks_tropical.isel(tracks=(
    (dates.year == 2020) & 
    (dates.month == 6) & 
    (dstracks_tropical.track_duration > 20) & 
    (dstracks_tropical.track_duration < 40)
)).isel(tracks=itrack)  # just select one track that meets criteria.
duration = track.track_duration.values.item()
duration

26

# Create a figure with 1 row and 2 columns
fig = plt.figure(figsize=(16, 6))

# Create left subplot (regular axis)
ax1 = fig.add_subplot(1, 2, 1)
track.ccs_area.plot(ax=ax1, label='CCS Area')
ax1b = ax1.twinx()
track.pf_area.isel(nmaxpf=0).plot(ax=ax1b, color='green', label='PF Area')
# ax1b.spines['right'].set_position(('outward', 60))  # Move the third y-axis to the right
# ax1b.invert_yaxis()
ax1b.set_title('')
ax1.legend()

# Create right subplot (with Cartopy projection)
# Use PlateCarree projection (equirectangular projection)
ax2 = fig.add_subplot(1, 2, 2, projection=ccrs.PlateCarree())

ax2.coastlines()
ax2.add_feature(cf.BORDERS, linewidth=0.4)

ax2.scatter(track.meanlon.values[0], track.meanlat.values[0], marker='o', color='k')
ax2.scatter(track.meanlon.values[duration - 1], track.meanlat.values[duration - 1], marker='x', color='darkorange')
ax2.plot(track.meanlon.values, track.meanlat.values)
add_ccs_area_swath(ax2, track)
buffer = 10
extent = track.meanlon.min().item() - buffer, track.meanlon.max().item() + buffer, track.meanlat.min().item() - buffer, track.meanlat.max().item() + buffer
ax2.set_extent(extent)

fig.tight_layout()
# fig.tight_layout(rect=[0, 0.03, 1, 0.95])  # Make room for suptitle

def convert_to_cftime(datetime, calendar):
    """
    Convert a pandas.Timestamp object to a cftime object based on the calendar type.

    Args:
        datetime: pandas.Timestamp
            Timestamp object to convert.
        calendar: str
            Calendar type.

    Returns:
        cftime object.
    """
    if calendar == 'noleap':
        return cftime.DatetimeNoLeap(datetime.year, datetime.month, datetime.day, datetime.hour, datetime.minute)
    elif calendar == 'gregorian':
        return cftime.DatetimeGregorian(datetime.year, datetime.month, datetime.day, datetime.hour, datetime.minute)
    elif calendar == 'proleptic_gregorian':
        return cftime.DatetimeProlepticGregorian(datetime.year, datetime.month, datetime.day, datetime.hour, datetime.minute)
    elif calendar == 'standard':
        return cftime.DatetimeGregorian(datetime.year, datetime.month, datetime.day, datetime.hour, datetime.minute)
    elif calendar == '360_day':
        return cftime.Datetime360Day(datetime.year, datetime.month, datetime.day, datetime.hour, datetime.minute)
    else:
        raise ValueError(f"Unsupported calendar type: {calendar}")

# These can be used to work out which pixel-level data to locate the times.
calendar = ds['time'].dt.calendar

base_times = track.base_time.values[:duration]
track_dates = [convert_to_cftime(pd.Timestamp(d).to_pydatetime(), calendar) for d in base_times]
print(len(track_dates))
track_dates

26

[cftime.DatetimeNoLeap(2020, 6, 1, 0, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 1, 1, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 1, 2, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 1, 3, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 1, 4, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 1, 5, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 1, 6, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 1, 7, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 1, 8, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 1, 9, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 1, 10, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 1, 11, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 1, 12, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 1, 13, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 1, 14, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 1, 15, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 1, 16, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 1, 17, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 1, 18, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 1, 19, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 1, 20, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 1, 21, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 1, 22, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 1, 23, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 2, 0, 0, 0, 0, has_year_zero=True),
 cftime.DatetimeNoLeap(2020, 6, 2, 1, 0, 0, 0, has_year_zero=True)]

# Track number in the mask file is +1 offset to the track index in the stats file.
track_number = track.tracks + 1

# Increase the animation embed limit [MB]
mpl.rcParams['animation.embed_limit'] = 30

%%capture
# Prev line ensures figure not shown until animation.fig, ax = plt.subplots(subplot_kw={'projection': ccrs.PlateCarree()})
# Set up a figure to use for the animation below.

margin_degree = 10
minlon = track.meanlon.values[:duration].min()
maxlon = track.meanlon.values[:duration].max()
minlat = track.meanlat.values[:duration].min()
maxlat = track.meanlat.values[:duration].max()
aspect = (maxlon - minlon + 2 * margin_degree) / (maxlat - minlat + 2 * margin_degree)
fig, ax = plt.subplots(subplot_kw={'projection': ccrs.PlateCarree()})
fig.set_size_inches((10, 10 / aspect))
fig.subplots_adjust(left=0.05, right=0.95, bottom=0.05, top=0.95)

# Create an animation of the track, MCS, OLR, and precip.
def plot_track_link_pixel(i):
    try:
        print(f'{i + 1}/{duration}')
        # print(f'{i}/{duration}')
        ax.clear()

        ax.set_extent((minlon - margin_degree, maxlon + margin_degree, minlat - margin_degree, maxlat + margin_degree))
        ax.coastlines(zorder=3)
        ax.add_feature(cf.BORDERS, linewidth=0.4, zorder=3)

        date = track_dates[i]
        ax.set_title(date)
        # Subset the HEALPix dataset to the current date & time
        _ds = ds.sel(time=track_dates[i], method='nearest')

        # Colormaps
        cmap_pr = mpl.colormaps.get_cmap('jet')
        cmap_rlut = mpl.colormaps.get_cmap('Greys')
        levels_pr = [0.5,1,2,4,8,16,32,64]
        norm = mpl.colors.BoundaryNorm(boundaries=levels_pr, ncolors=cmap_pr.N)
        # Precipitation
        _pr = _ds.pr * 3600000  # Convert unit to mm/hr
        _pr = _pr.where(_pr >= np.min(levels_pr), np.nan)
        # Mask
        _mask = _ds.mcs_mask.load()
        _mask = _mask.where(_mask == track_number, 1, 0)
        Zm_mask = np.ma.masked_where(_mask == 0, _mask)
        # Colorfill OLR and precipitation
        im_rlut = egh.healpix_show(_ds.rlut, ax=ax, vmin=80, vmax=260, cmap=cmap_rlut, alpha=0.7, zorder=1)
        im_pr = egh.healpix_show(_pr, ax=ax, cmap=cmap_pr, norm=norm, alpha=0.9, zorder=1)
        im_mask = egh.healpix_show(Zm_mask, ax=ax, cmap='Reds', vmin=1, vmax=1.01, alpha=0.5, zorder=2)
        # Contour MCS mask (boundary)
        # im_mask = egh.healpix_contour(_mask, ax=ax, levels=[0.5], colors=['r'], linewidths=1, alpha=1)

        # Display track path.
        ax.scatter(track.meanlon.values[0], track.meanlat.values[0], marker='^', c='maroon', zorder=3)
        ax.scatter(track.meanlon.values[duration - 1], track.meanlat.values[duration - 1], marker='x', c='maroon', zorder=3)
        ax.scatter(track.meanlon.values[i], track.meanlat.values[i], marker='o', c='firebrick', zorder=3)
        ax.plot(track.meanlon.values, track.meanlat.values, 'r-', zorder=2)
        # clear_output(wait=True)

        return ax
    except Exception as e:
        print(f"Error in frame {i}: {str(e)}")

# Create animation with specific settings
anim = matplotlib.animation.FuncAnimation(
    fig, 
    plot_track_link_pixel, 
    frames=duration,  # Number of frames
    # frames=2,
    interval=500,
    blit=False,  # Set to False for complex plots with multiple elements
    cache_frame_data=False  # Disable caching to avoid memory issues
)
anim

1/26
1/26
2/26
3/26
4/26
5/26
6/26
7/26
8/26
9/26
10/26
11/26
12/26
13/26
14/26
15/26
16/26
17/26
18/26
19/26
20/26
21/26
22/26
23/26
24/26
25/26
26/26

Once Loop Reflect

• Track: red line with circle showing its centroid.
• MCS mask: red shading.
• OLR: grey filled.
• precipitation: color filled contours.